CENTER FOR ENVIRONMENTAL STRUCTURE

HOUSES GENERATED
BY PATTERNS

CHRISTOPHER ALEXANDER

SANFORD HIRSHEN

SARA ISHIKAWA

CHRISTIE COFFIN

SHLOMO ANGEL
 ACKNOWLEDGEMENTS

Many friends helped us with this book. We

should especially like to thank Emiliano Aviia,
Antonio Pinto, Alvarado Villegas and Pablo Eugenio
Roman, who took us into their families while we
were in Lima. Without their help we could never
have understood enough about the Peruvian way of
life. Catalina de Martin wrote the Spanish sum-
mary of the book, which is available separately.
Gary Saxton made the perspective sketches; Bill
Grindley helped to criticise our assumptions about
Peruvian culture; Ron Gammill provided design input
and criticism and helped with drawings; Steve
Gilmore built, tested and revised structural com-
ponents; James Brown and Larry Kaltman of Technol-
ogy Consortium Inc., gave us great quantities of
time and equipment to help our beam and plank
tests; Hector Gallegos and Carlos Li Carillo gave
us detailed cost data for Lima; Barbara-Katharina
Schreiner edited the book; Mary Louise Rogers typed
it; and Shula Angel and Maura Kealey proofed it.
Denny Abrams and Van King helped in the panic of
production, and Bill Buckman, Cece Weston, David
Goines and Bret Buckman worked day and night to
print it in time.

Finally, we should like to thank Peter Land

for having created the whole idea of the Proyecto
Experimental de Vivienda and the competition, which
gave us the chance to make these designs.

Copyright © 1969 by the Center for Environmental Structure

2531 Etna Street, Berkeley, California

Second Printing 1970

a
CONTENTS

THE DESIGNS
7

THE PATTERNS
51

-3-
PREFACE: THE PROYECTO
EXPERIMENTAL DE VIVI ENDA

In January 1969 the United Nations, working with the

Banco de la Vivienda of Peru, asked us and twelve other archi-

tects from various countries to submit competition designs for

a community of 1500 houses. The houses are to be built at a

gross density of 37 houses per hectare on a site of 40 hectares,

8 kilometers north of Lima. The site is bounded by two major

arterial highways, and crossed by a third; these highways are

fixed by the Lima transportation plan. The financial arrange-

ments require that each house be contractor built, on its own

land, at a cost ranging from 78,000 to 164,000 soles ($1800 -

$3800), and that they be sold to low income white collar work-

ers (empleados) earning between 2,800 and 5,800 soles ($65 -

$135) per month.

We were asked to present our designs in a way that would

help the evolution of Peruvian community and house design in

the future. We have . therefore chosen to present our work in

two parts:

In part one we present our designs. This includes a site

plan, drawings and construction details for individual houses,

-5-
and a choice process which allows the final site and house

plans to be formed, in detail, by the idiosyncratic needs of

the individual families who buy the houses.

In part two we present sixty-seven general design

principles which we call patterns. These patterns describe, in

an abstract sense, the lessons which a Peruvian architect might

learn from our designs, and could re-use, over and over again,

in his own designs . We have combined these sixty-seven patterns

in one particular way, to form the designs presented in part one.

However, in the hands of different Peruvian architects, or in

the hands of different individuals designing and building their

own houses, these patterns can generate an almost infinitely

rich variety. In this sense, these patterns -may begin to define

a new indigenous architecture for Peru.

-6-
THE DESIGNS

Drawings

The Generic House 8

The Site 10
Houses on the Site 12
Construction 14
The Site 16

The House 20

Construction 24

The Choice Process 29

The Combination Process 38

Costs 48

1\.
 Second Floor

First Floor

0 1 2 3 4 5 Meters

1 Entrance 9 Kitchen Patio

2 Sala (Parl or) 10 Storage Patio
3 Family Room 11 !'1aster Bedrcqm
4 Family Room Alcove 12 Bed Alcoves
5 Main Patio 13 Mirador
6 Veranda 14 Clctl-.es uryina Clcset
7 Kitchen 15 S how er
8 Laundry 16 Toilet

THE GENERl .C
 1 Primary School 9 Grove of Trees
2 Secondary School 10 Kindergarten
3 Technical Secondary School 11 Clinic

sIT E
4 Church 12 Dance Hall

North 0 50 100 200 Meters

5
6
7
Cinema
Supermarket
Market
13
14
15
Sports Center
Parking
Outdoor Room
THE
+ 8 Municipal Offices

I
 1

I
I
.·1. l

~·.:,:.:,: /.

. ·;::. ;_.

/t"'""''>~::~J
.., ;
..

::
':

·:. ·.-····

' I
I
CONSTRUCTION \
I
I
1 Floating Slab

· . . n\\ l \
2
3
4
Bamboo - Urethane Foam Beam
Bamboo - Urethane Fo~ Plank
Mortarless Block Cavity Wall
\
THE SITE

The site contains 1726 houses , at a gross density of 43

houses per hectare. House lots are 5 . 20 meters wide , and vary

in length from 13 to 27 meters .

No two houses are alike . The exact form and length of

each house is determined by a choice process which allows f am-

ilies to fit their houses to their own needs and budgets .

(This is fully described on pages 29-37 . )

Since the lengths of houses in the final site plan will

be based on the choices which families make , and are unknown at

present, the current drawing of the site plan is only approx i-

mate. Once each family has made its choices , it will be neces-

sary to lay out a new site plan. This new plan will have the

same morphology as the one shown, but the exact number of

houses of different lengths, will reflect the families ' choices .

The morphology of the plan is fluid enough to adjust to the new

lengths .

The site contains a number of cells . Each cell contains

30-70 houses; it is a pedestrian island, surrounded by a sunken

one lane road, which feeds small parking lots that surround the

cell.

We have designed the cells with the - idea

- - -- - - - -
that the particu-

lar group of people who live in a cell can make an impact on

-16-
 their cell, can give it a unique atmosphere, created by them,
-~----

'-and can then , in a real sense "take possession of it".

First , the basic form and circulation of each cell is

unique - according to its particular location in the large plan.

Second, during the choice process, people will be asked

questions about the location they want for their house . When

they are then located according t o these choices, people with

similar attitudes and interests will be living in the same cell .

Third , the cells are physically separated , and the pedes-

trian passes through a physical gateway whenever he enters a

cell : this will give each cel l a better chance to build up its

own unique flavor .

Fourth , at the heart of each cell~ there is a small open

place , surrounded by an unfinished, roofed arcade . It is our

intention that the people who live in the cell will develop

this arcade according to those community uses they think most

valuable. Detailed discussion

of the cell concept, and of the

interior organization of the

cells, is given on pages 55-61

and 99-112.
Over and above the cells ,

the site contains three major

overlapping configurations: t h e

road system, the pedestrian net-

work, and the community spine .

-17-
 Vehicles travel on narrow one-way loop roads, around the

cells, with car parking at the entrances to the cells. There

are enough parking spaces to provide for 50% car ownership.

This figure was given to us by the United Nations: they esti-

mate 50% car ownership in 30 years, and asked us to work to

that figure. For details of the road system, see pages 63-70.

The central spine of the pedestrian system, we call the

"paseo". This name is taken from the Latin American habit of

the evening and Sunday stroll (paseo in Spanish) . The paseo

gives people a high density pedestrian spine of looped paths

where a tradition of evening and Sunday walks . can develop.

At frequent intervals along the paseo, there are "activity

nuclei": small open places, with the community facilities and

shops grouped around them. The community facilities are grouped

in such a way as to create a special character at each of these

activity nuclei. For details of the community facilities, see

pages 87-97.

-18-
 The peripheral pedestrian paths connect cells to one

another, and connect them to this main paseo. Each cell which

is large enough, has a pedestrian loop· in it: this will help

to create the inner character of the cells, since it will be-

come natural for people to take a walk "around the cell". All

pedestrian paths from the outer parts of the site , lead towards

one of the eight activity nuclei : The nuclei will always be

full of people . For the pedestrian network, see pages 71-86 .

-19-
THE HOUSE

Although the choice process guarantees that no two houses

will be exactly alike, all houses are based on one generic house.

This generic house is a two storey house, 5.20 meters wide,

and about 20 meters long, which has an alternation of rooms and

patios along its length, the rooms connected by deep verandas.

This alternation gives every room light and air, and makes the

house seem larger. The two main patios are always one behind

the other in the direction of the breeze (which comes from the

south) - so that cool air circulates through the house in sum-

mer. In winter, the patios will be covered by dacron sailcloth

covers which run horizontally on rods at roof level. They may

be controlled from upstairs by cords, and make the patios usable

all the year round. Patios are discussed fully on pages 113-123
and 181-190.

-20-
 The ground floor of the house contains two parts: a

public part and a family part. The main features of the public

part are the front patio, and the sala (formal living room or

parlor). In Peruvian life there is a strong distinction between

members of the family, who may go anywhere in the house, and

strangers, who must be entertained in the sala. The sala is

rr
L: • - •
n

separated ·from the rest of the house by the front patio, and,

even in the smallest house allows visitors to be treated with

proper formality. Detailed discussion of the public part of

the house is on pages 127-138.

The family part of the house centers around the family

room (comedor estar) . An alcove (two in large houses) opens

off this family room to make a place where children study at

night, where women can sew, where people can talk while the

TV is on, etc. Behind the family room there is a kitchen, with

two service patios, one on either side of it. The one between

-21-
kitchen and family room is a pleasant place, where people can

.eat, and work. The other provides storage for the inevitable

building materials, animals, and laundry lines. For details of

the family pa~t of the house see pages 139-151.

Upstairs the house contains a master bedroom, bathroom,

and a number of tiny individual bed alcoves. These bed alcoves

give each child a small space which is his own, for his own

things; very young children may double bunk in a single alcove.

Since Peruvians don't like being isolated, these alcoves are

clustered around common spaces. There are two clusters; one for

boys and one for girls . Every house , even the smallest , can be

extended to make room for as many as eight beds . For detailed

discussion of the sleeping areas, see pages 153-165 .

Every house can be extended to provide room for a grand-

mother downstairs near the family room , a sleeping place for a

servant , a room at the back which can be rented out , or a smal l

shop . Many low income fami l ies try to mak e ext ra mone y by rent-

ing out rooms or selling things : These e x tensions a r e much

easier if the b ack of the house opens onto a pedes t ria n way :

those people who are ~illing to pay fo r i t , may choose this

-22-
 option in the choice process. For details of these extensions,

see pages 191-194.

Finally, each house has a very strongly marked entrance,

with deep recesses, a seat outside, and a gallery or "mirador"

at the second storey. Peruvians spend a great deal of time

street watching: people hang out in doorways, sit on benches

outside the doors, and watch the street from windows above.

They like to be in touch with the street, but from the seclu-

sion of their homes. Most houses in our site plan command a

direct view into the center of the cell in which they stand,

so that activity can be seen from the front window or the door.

Details of the entrance are on pages 167-180.

- ------

I -23-
CONSTRUCTION

The basic structure of the house consists of a floating

slab foundation, load bearing walls, and a light weight plank

and beam system . This form of construction is conceptually

very similar to traditional construction : but each of the com-

ponents is a cheaper , lower-weight higher-strength version of

its traditional equivalent . The floating slab is laid in large

sections by a road building machine. The walls are interlock-

ing mortarl~ss concrete-block walls, reinforced with sulphur ,

with a cavity for plumbing and conduits . The planks and beams

are made of urethane foam-plastic and bamboo, reinforced with

a sulphur-sand ·topping .

/r. ·"I' . • .

--. -- --.
!"'-- -- .
,
+i - ... •
-~ '

Experiments
 All these building components can be produced in Peru

today with available resources and skills. Further, the ideas

embodied in these methods and products have the potential for

long-range development of natural resources. These, in turn,

will directly contribute to the economic growth of the people

and the country, a vital factor in creating a national housing

policy. Sulphur is available in huge quantities in Peru; cur-

rent estimates show 50,000,000 tons of sulphur waiting to be

mined in the Peruvian Andes. The use of urethane foams has

been tested in various parts of Latin America; it is a seed

industry of great importance, since foams are now used in many

different ways, inside and outside the building industry. A

urethane plant, once started , would benefit many sectors of

the national economy . Bamboo is widely available in the north

of Peru, and may be imported cheaply from Ecuador. Though it

is often thought of as a low-prestige material, it will quickly

become a material of great value when used together with high

performance bonding agents.

These building materials are especially suited to the

local earthquake conditions. The mortarless block has been

tested in Mexico, under earthquake conditions, and has per-

formed well throughout. Sulphur is now being tested in earth-

quake zones, and performs as well as steel reinforcing. The

floating slab has a long history of success in earthquakes.

The urethane-bamboo sandwich is enormously strong - and its

-25-
 )

very light weight reduces loads during earthquakes. Technical

details and performance data for all building components are

given on ~ages 195-219.

These materials are not merely low cost, high performance

materials. The yellow diamonds of the block wall, where it is

reinforced with spots of sulphur, the warm texture of the bam-

boo ceilings, the deep polished red of the oil stained concrete

slab, and the translucent white of the dacron sailcloth patio

cover , combine to create a house which is far more warm and

human than the usual heavy grey of low cost construction.

To simplify building construction, all components are

prefabricated, on site. They all conform to the 10 cm module.

They are assembled dry. This makes them equally suitable for

use by the contractor, when the houses are first built, and by

the families who live there, when they want to change their

houses later.

We have chosen these components with special emphasis on

the idea of future do-it-yourself construction. Peruvian fami-

lies add to their houses, and change them, continually. They

can only do this if the components are e~tremely small in scale,

and easy to work with home tools. We have therefore tried very

hard to create a system of components that are easy to work,

and can be used at the rather low tolerances that correspond

to the realities of home construction. In our opinion, this

is more relevant to people's needs, than a system of highly

-26-
machined components, which must be built to very fine tolerances.

Given the assumption that home construction will always be done

rather roughly, with hammer and nails, and fillers where re-

quired, our system will allow the homeowner to do almost any-

thing he wants to do.

For example:

On the slab foundation, a new wall can be built anywhere,

without needing extra footings . The mortarless block wall can

have individual blocks removed or added, at will . The hollow

wall makes it easy to add new plumbing fixtures or electrical

conduit , cheaply and simply, by taking out a block . A person

can make his own blocks, instead of buying them: the block

moulds are designed to be operated by one unskilled person .

Extra block columns can be inserted at any point . The sulphur

joints , unlike cemented joints , need only to be melted by l ocal

application of heat , to l oosen ; when they cool they harden again .

The bamboo foam beams are made in five meter lengths which fit

across every house ; they can sit anywhere along the length , on

the continuous impos t b l ock . They can be hand cu t to frame any

desired opening . The bamboo foam p l anks can a l so be h a nd cut

to any l ength and any width. The beams which suppor t t he roof

a r e i nit i a l l y designed to c arry a minimal live load on l y : if

the house own er wants to mak e a usab l e third storey , he may i n-

sert e xtra b eams next t o t he e xi sting ones .

F i na lly , the component s mu s t be easy t o get . They wi ll

be impossibl e t o ge t o n t he open marke t . To make them avai l abl e ,

-27-
I
we propose that the community contain a new kind of community

service, which we call the Community Building Supermarket. This

supermarket will start as the on-site factory needed for initial

construction. In its later life, it will manufacture components,

sell them, rent out the equipment needed for assembly, provide

skilled labor for those aspects of the construction which in-

volve new techniques, and train members of the community who

want to learn these techniques for themselves.

-28-
THE CHOICE PROCESS

The people who live in our houses will, because they are

all Peruvians, share certain needs and all have similar back-
~:y,;,\ :..,-Y).,< •
grounds . At the same time, each person, and each family, will

be unique. The choice process tries to do justice to this fact .

The needs which people share led us to the patterns de-

fined on pages 51 - 219 and these patterns led us to the generic

house design already presented . But even if all families share

the needs which are solved by this generic house, they will , be-

-cause they are unique, also have very different attitudes t o

the _relative importance of these different needs . One family,

which tends to be formal, will consider the need for a sala

most important of all; another family in which life tends to be

informal, may live most of the time in the kitchen. Although

both families will want a sala and a kitchen, the first family

would prefer a large sala and small kitchen, if they had to

choose - and the second family a large kitchen and a small sala.

The uniqueness of any" family, will, in this way, be reflected

by the relative amounts of money they would want to spend on

satisfying their various needs.

~his is essentially how the choice piocess works. We

ask each family to decide how much they want to spend; and then

-29-
 )

we ask them to divide this money up among the various parts of

the house, in the way that best reflects their individual pref-

erences. ' The form of the house allows its various parts to vary

in size, independently of one another; without disturbing the

unity of the whole .

Even though no one part of the house can take more than a

small number of different siz~s, the total number of combinations

lU

---~
~/

•' r::-
t __ .;

•I - _, · - _:'.i t,_,,_
~- I i i i , ' I ; • · ·
,.._..,.r:,.~ - ·~·
I ; _r ; •

:::z ' r--,

i I '
is extremely large - in the neighborhood of a million. In a

community of 1500 houses, it is highly unlikely that any two

will be the same. This variety is not just visual variety: it

is variety which reflects the real variety of attitudes to life

which will exist among the fifteen hundred families who live in

the Proyecto Experimental.

On the following pages we present the choices which a

family would have to make, before buying a house in the Proyecto.

These choices would, of course, have to be made before construc-

tion starts. To help people make the choices, it would be essen-

tial to build one or two model houses ahead of time, and allow

people to visit them. Otherwise they would probably not be able

to grasp the meaning of the choices.

The combination process, which translates the completed

choice sheets into designs, is presented on pages 38-47.

-31-
 PROYECTO EXPERIMENTAL DE VIVIENDA

FAMILY CHOICE SHEET

By answering the questions on this sheet, you may decide

how big to make the various rooms inside your house. Start by
deciding how much money you want to spend on the house, alto-
gether. You may choose any amount between 79,000 soles and
163,000 soles.

When you have decided how much you want to spend a ltogether ,
you may start making choices about individual rooms and finishes,
and about the location of the house in the community. Each of
these choices costs money. To make your choices, you wi ll have
to decide which things m~tter most to you.

The available choices are shown below. The numbers which

follow each item, show, in thousands of soles , how much each of
the various choices costs. Write the number you have chosen
for each item, in the column on the right. When you have fin-
ished, add up the numbers on the right. You must be sure that
the total is the same as the total price you are willing to pay
for your house. Thus, for instance, if you want to spend 95,000
soles on your house, the numbers in the rig ht hand column must
add up to 95.

On the pages which follow this one, we give you the de-
tailed explanation of these choices one by one, so that you can
understand clearly what you are choosing. You should read these
detailed explanations before you try to make your choices.

Your
Choice
SALA Choose one of these 3 8 14 20

FAMILY ROOM Choose one of these 22 24 26 28 32 35

MAIN PATIO Choose one of these 18 21 23 27

KITCHEN - LAUNDRY Choose one of these 13 22

BED ALCOVES Choose o ne of these 6 9 13 16 19 22 '0

MASTER BEDROOM You must choose this 11

GRANDMOTHERS BED ALCOVE Choose one of these 0 1

LAUNDRY-STORAGE PATIO Choose one of these 6 7 8 10

CAR HOUSE DISTANCE Choose one of these 0 3

RENTAL/BACK DOOR Choose one of these 0 2 4 8 10

SHOP Choose one of these 0 8 10 20

EXTRAS You may choose more than one 1 1 1 1 2 3 5

of these, or none. Write
the total on the right

FINISHES You may choose more than 1 1 2

one of these , or none .
Write the total on the
right
TOTAL :

-32-
 PROYECTO EXPERIMENTAL DE VIVIENDA
LIMA, ,EAU

l. SIZE OF SALA

In all houses, the sala is a receiving room, separate from

the comedor-estar. The very smallest sala you can have is a
tiny alcove just large enough to hold two chairs and a sofa,
inside the front door. Or you can have a sala more like an
ordinary room.

Choose one:

Tiny sala COST: 3,000 soles

Tiny sala with option for medium
sala later COST: 8,000
Medium sala COST: 14,000
Large sala COST: 20,000

2. SIZE OF FAMILY ROOM

Your family room can have a number of different sizes. All

of these rooms have one small alcove off them, where children
might do their homework, or a woman might sit and sew. In the
smallest one, this alcove is very small. In the larger ones,
there is a second alcove.

Choose one:

Small room + one (small} alcove COST: 22,000 soles

Small room + one alcove COST: 24,000
Medium room + one alcove COST: 26,000
Large room + one alcove COST: 28,000
Mediwn room + two alcoves COST: 32,000
Large room + two alcoves COST: 35,000

As you will see on the next page, the master bedroom size
is determined by the family room size which you choose. For
this reason the costs of the family room shown here include
the extra costs created by larger master bedrooms.

3. MAIN PATIO

Every house has a main patio, which contains the stair, is

flanked by the sala, the family room, the entrance, and by a
veranda which connects the sala and family room . This patio is
covered in winter by a dacron sailcloth. The patio is always 3
meters wide, but its length may vary. Especially if you intend
to plant things there, you should choose one of the larger ones.

Choose one:

Tiny patio COST: 18,000 soles

Small patio COST: 21,000
Medium patio COST: 23,000
Large patio COST: 27,000

-33-
 PROYECTO EXPERIMENTAL DE VIVIENDA

llMA, "E Pl U

4. KITCHEN-LAUNDRY

There are two kitchens to choose from. The small one has
no sitting space in it , and ha s the laundry counter out of doors,
in the storage patio . The l arge one is twice as large, has room
for a table in it, and has the laundry counter under cover. Both
kitchen s contain, a s an integral part, a small kitchen patio,
whi ch wi ll be covered in winter, and gives you extr a room to work,
room for childre n to play, e tc.

Choose one:

Small kitchen COST: 13 , 000 soles

Large kitchen COST: 22,000

5. NUMBER OF BED ALCOVES

Every house will have a master bedroom; but th is is the only

conventiona l bedroom. All the o t her beds will be in individual
bed alcoves, containing a bed , a dressing space, and sto rage ,
and nothing more. These be d alcoves will be arranged in two
small c lust ers , one for boys and one for girls .

At least one of the alcoves will be an extra large one , large

enough to hold two bunk beds. If you would lik e to have two of
your children sleeping in bunk beds li ke this, you ma y choo se a
number of alcoves which is one l ess than the number of child ren
in your fami l y .

You may choose how many individual bed alcoves you want (e?.ch
alcove costs 3,000 so l es) , and you may also sa y how you want the
alcoves clustered. Choose one of the following. (The fi r s t num-
ber is the number in the front cluster, the s e cond in the back.)

Two beds: 2,0 COST: 6,000 soles

Three beds : 3' 0 or 2' 1 COST: 9,000
Four beds : 4' 0 or 3,1 or 2,2 COST: 13,000
Five beds: 4' 1 or 3,2 or 2,3 COST : 16,000
Six beds: 4' 2 or 3 , 3 or 2,4 COST : 19,000
Seven beds: 4' 3 COST : 22 , 000

6. MASTER BEDROOM , BATH AND STORAGE

The master bedroom and s t orage are always above the family
room . These r ooms can vary in size, but the i r size must always
correspond to the s ize of the family room you have chosen . In
this sense, you have no r ea l choice of sizes here, though it may
in f luence your choice of family room size : the bathrooms are al-
ways the same size.

The fixed cost of this room is 11,000. All additional costs

are included in the costs of family rooms. For all family rooms
the master bed costs are:

Master bedro'om COST: 11,000 soles

-34-
 PROYECTO EXPERIMENTAL OE VIVIENOA

LIMA, "E"U

7. GRANDMOTHER'S BED ALCOVE

Every house has the possibility of placing a bed alcove, on

the ground floor, just next to the family room, for an old person.
You may choo s e to have this bed alcove built today, or you may
build it yourself, later.

Choose one:

Grandmother's bed alcove built now COST: 1,000 soles

Not built now COST: zero

8. LAUNDRY-STORAGE PATIO

Every ho u se has, at the very back, an extra patio which gives

you room for expa nsion on the ground floor, a place where se rvant
quarters may be constructed, a place to hang laundry, and to stor e
building materials and large objects. If you want to u se this
patio to build rental space, or a little shop, do not choose the
very sma ll patio.

Choose one:

Very small COST: 6,000 soles

Sma ll COST: 7,000
Medium COST: 8 ,000
Large COST: 10,000

9. CAR HOUSE DISTANCE

Some houses have parking lots ri gh t n ext to them, others do

not . In no case is a house more than 50 meters from the nearest
parking lot. If you have a car or if you expect to have a car
in the near future, you may want to choose a house near a parking
lot . Otherwise you may prefer the quietness and safety of a
pede strian stree t, especially if you have children who play out-
side the house.

Choose one:

There is a parking place within 15

meters of your house COST: 3,000 soles
Ther e is a parking place between 15
meters and 50 meters from your
house COST: zero

10. RENTAL AND/OR BACK DOOR

If you hope to rent out a small room in the future, or if

you are particularly anxious to let your servant have a back
door which is separate from the front door, you may have a lot
which ha s a back or a side opening onto a walkway. If you choose
a lot that has a second entrance , you may have a small room for
rental, built there today, or you may leave it unbuilt, and build
it yourself later. If you choose a shop you may not choose a
rental unit.

-35-
 PROYECTO EXPERIMENTAL DE VIVIENDA
LIMA. !'£RU

Choose one:

Corner lot with second entrance on

side and rental space built today COST: 10,000 soles
Corner lot with second entrance on
side COST: 8,000
Lot with second entrance on back
and rental space built today COST: 4,000
Lot with second entrance on back COST: 2,000
Lot with front only COST: zero

11 . SHOP

You may want, now or in the future, to open a small shop.

In this case you will want to have a location where a shop can
prosper - either next to the main market, or at some corner
which many people are going past. If you choose a shop loca-
tion by the market, you will have the shop built now, auto-
matically. If you choose one of the other shop locations, you
ma~ have the shop built now, or you may choose to leave it un-
built now, and then build it yourself later. If you have chosen
a rental space, you may not choose a shop.

Choose one:

Market location, with shop built now COST: 20,000 soles

Corner location, with shop built now COST: 10,000
Corner location, shop not built now COST: 8,000
No possibility of building a shop COST: zero

12. EXTRAS

If you find that you can purchase adequate space for your
family without using all the money that you plan to spend on
housing, or if some of these extra features mean more to you
than extra space, you may want to choose some of the optional
features listed below.

Fiberglass patio covers instead of

sailcloth on front patio and middle
patio COST: 5,000 soles
Electric hot water heater, connected
to bathroom, kitchen and laundry COST: 3,000
Second wa s h basin for family COST : 1,000
Wash basin and WC for servant,
e nclosed COST: 2,000
Colored tiles around the main door COST: 1,000
Bench near the main door COST: 1,000
5 meter Eucalyptussaplings near
front door COST: 1,000

-36-
 PROYECTO EXPERIMENTAL DE VIVIENDA
LIMA, ltERU

13. FINISHES

In Proyecto Experimental the basic finishes are the

finishes of the building materials, block and bamboo. If
you want to, you can have plaster or white wash in the sala
or comedor.

Plastered ceiling in the sala COST: 1,000 soles

Plastered ceiling in the comedor-
estar COST: 2,000
White washed block walls COST: 1,000

Finally, you may choose where you want your house to be

in the larger community. There are two choices: neither has
any cost.

14. QUIET AREA OR BUSY AREA

There are two kinds of housing area to live in. In one

kind of area there are many people walking up and down the
pedestrian way outside the houses. During the day and in the
evening, there will almost always be some people on the street.
In the other kind of area there are far less people walking
through .

Choose one:

Many people going past your house.

Few people going past your house.

15. NEARBY COMMUNITY FACILITIES

The church, market, clinic, parks, kindergartens, schools,

secondary school, technical school, evening entertainment, are
all fixed positions on the site, and each has a housing area
that is particularly closely tied to it. Choose the one which
you would most like to be near:

Church
Market
Clinic
Park
Kindergarten
Schools
High school
Technical school
Evening entertainment

* * *

-37-
THE COMBINATION PROCESS

The combination process is not unlike the process by

which the l~~,i~ s •on a tree are formed. All the leaves are de-

fined by the same morphogenetic rules: the individual leaves

are formed by the interaction between these rules and the lo-

cal conditions which the leaves are subject to. As a result,

each leaf turns out unique, according to its position in the

whole tree; yet in a generic sense the leaves on the tree are

all the same. The combination process works in the same way.

All houses are formed by the same sequence of rules,

based on the form of the generic house. But each house has to

me et certain particular conditions: those imposed on it by the

famil y 's choices, and those imposed on it by its position in the

site - orientation, the lengths of next door houses, location of

nearby pedestrian walkways, and so on. Each individual house is

formed by the intera-ction of _the _local conditions which it has

_t,o meet, and the g~ neric r_ule _s of the combination process.

For example, in order to make the house-form coherent,

the shape of the house entrance must be different for houses

with a small sala and houses with a medium sala, it must be

different for houses on a corner lot and houses on a center lot,

it must be different according to the length of the next door

-38-
house on the eastern side (since the entrance is always on the

east). The rules which form the house entrance (Steps 6, 7, 8

f .
below), therefore depend on the size of the sala, the type of

lot, and the position of the next door house on the eastern side.

It is only this high degree of interaction between the

rules of the combination process and the local conditions which

guarantees that all houses are internally coherent, and that

each house fits coherently into the larger site plan. Like the

leaves on a tree, all the houses will be different, yet all of

them coherent and all of them the same.

It is very important to stress the fact that the rules of

the combination process are almost mechanical, and can be carried

out by any trained draftsman. The low cost of the houses cannot

support any individual design time. We estimate that a trained

draftsman will need about one hour per house, to translate the

family choice sheet into a set of working drawings and specifica-

tions for thel contractor.

The draftsman has one master site plan with the house lots

shown on it: and one file for each family, containing the family

choice sheet, and a blank house plan, which shows the side walls

only, 5.20 meters apart, for both floors, and shows no end walls

or interior walls. He now builds up the detailed design of each

house, by using the following rules, one step at a time:

-39-
STEP 1: Assignment of houses to cells

Assign each family to a cell in the site plan, on the basis

of their answers to questions 14 or 15 on the choice sheet.
Location across the site is determined by choice 14. If
they want to be in a busy area, place them along the paseo.
If they want to be in a quiet area, place them far from the
paseo. Location along the length of the site, is determined
by the community facility they want to be near (choice 15).

STEP 2: Determination of house length

Fix the house length as the sum of the lengths of the chosen
sala, patio, family room, kitchen and back patio (choices 1,
2, 3, 4, 8).

STEP 3: Assignment of houses to sites, within the cell

,.\\\\ - '

Within the cell fixed by step 1, assign each house to a lot

whose length is as near as possible to the length determined
by step 2, and which also satisfies the family ' s choices con-
cerning shop location, rental/back door, and distance from
parking lots (choices 9, 10, 11). (At this point the new site
plan will be slightly different from the current site plan -
since each house will have a slightly different length. It
will now be necessary to make minor changes in layout, and
arrangement of houses, so that pedestri~n paths, loop roads
and parking lots still have a coherent \'.form.)

STEP 4: Detailed site conditions for each house

Since the house is now fixed within the site plan, the positions
of next door houses, positions of adjacent paths and roads, the
front end of the house and orientation of the house are now
fixed. Transfer these to the drawing of the individual house.

-40-
STEP 5: Position of patio
openings

The patio openings are always '

on the east side of the house. '--- ----

Sketch in patios for different

orientations as shown.

North

STEP 6: General position of

front door west Side Neither Side East Side
Ope n open Open

The position of the front door

is given as a function of patio
position (step 5) , and the
presence or absence of adjacent
houses . Sketch entry arrows
according to this table.

South
Facing

North
Facing

The arrangements for houses

facing north are mirror images
of the arrangements for houses
facing south. To avoid dupli-
cation , all future diagrams will
be shown for south facing houses
only. Arrangements for north
facing houses are obtained by North

taking the mirror images of

these arrangements .

-41-
STEP 7: Sala size and position Front Door Front Door. Front Door
To West To East
and exact position of front door At Front

The design of sala and front

door, in relation to one another,
are given as a function of door Tiny Sala

position (step 6) and the size

of sala shown on family's choice
sheet (choice 1). Draw sala and
entrance according to this table.
Tiny Sa l a
With Future
Medium Sal•

Medium Sala

Lill'.ge Sala

STEP 8: Detail of front door Medium or

Large Sala
Tiny Sala
if at front end

The front door is on the east House to East

Projects LL -L .~---l
___ J l
side of the front end of the
house (step 6) . Its detailed
treatment depends on the posi-
tion of the house next door and
to the east, and on the sala
House to East
Flual,l ll l U4.
size (step 7). Draw according
to this table. House to East
Behind {<2 meters) ll J- Li.+
House to East
Behind (>2 meters)
Li._} u:

-42-
STEP 9: Size of front patio MediWD. or
Tiny Sala Larqe Sala

The position of the front wall

of the family room alcove and
veranda column positions are
fixed according to the family IS Very Small Patio Not Poaaible

choice of patio size (choice 3),

and the size of the sala (step
7) . Draw according to table.
Small Patio mtu
Medium Patio
lI~ tD
Lar9e Patio
mtr1
STEP 10: Family room

The family room is fixed directly by the amount the family wants
to spend on it (choice 2). Draw as shown .

22,000 24,000 26,000 28,000 32,000 35,000

STEP 11 : Stairs Small Alcove

At Fron t
Lar9e Alcove
At Front
Of CO?lledor Of Comedor

The position and size of the

stair landing vary according to
the size of the front alcove .
Draw as shown.

-43-
-

Small Xitchen Small Xitchen

0 to l Beds in 2 to 3 beda in
Back Cluster Back Cluster Large Xitchen

STEP 12: Kitchen and laundry

ldld ~
The arrangement of the kitchen
is determined by the kitchen
size (choice 4), by the pres- No Back
Alcove in
ence or absence of a second Comedor

family room alcove (step 10),

~~ [tj ~
and by the number of beds in
the back cluster upstairs
(choice 5) . Draw according to
table. Back
Alcove in
Comedor

STEP 13: Master bedroom and bathroom layout

The layout of the master bedroom is given directly by the length

of the family room below (step 10). Draw as shown .

22,000 or
24,000 26,000 28,000 32,000 35,000

STEP 14: Back cluster of bed Small Kitchen Large Kitchen

alcoves

The layout of bed alcoves in the

back cluster depends on the num- No Bed Alcoves
ber chosen by the family (choice In Back

5), and the kitchen size (step

12). Draw according to table.

1 Bed Alcove
In Back

2 Bed Alcoves
In Back [] SS
-44-
J Bed Alcoves
In Back
Ea
Not Possible
STEP 15: Front cluster of bed alcoves

The layout of bed alcoves in the front cluster depends on the

number chosen by the family (choice 5). Draw according to
table.
Tiny Sala Tiny Sala Tiny Sala Medium Sala
Small Patio Medium Patio Large Patio very Sniall Patio

Two front bed alcoves

Medium Sala Medium Sala

small Patio Mediwn Patio

Medium Sala Mediwn Sala Medium Sala Medium Sala

Very Small Patio Small Patio Medium Patio Large Patio

Three front bed alcoves

&g ~ ~
Large Sala Large Sala Lar9e Sala Large Sala
Very Small Patio Small Patio Medium Patio Large Patio

Medium Sala Medium Sala Large Sala Large Sala

Medium Patio Large Patio Medi um Patio Large Patio

Four front bed alcoves

-45-
STEP 16: Mirador

Front mirador is given by step 15 . Side mirador occurs in

corner houses only. Draw according to table .

W••t Facinq Mirador ~~

Very Srnall
Patio Small Patio Mediwn Patio Large Patio

EaJJt Facing Mirador

STEP 17 : · Position of windscoop

The windscoop always faces the south wind. It is on the main

patio in a south-facing house, and on the kitchen patio in a
north-facing house. Draw in positions shown.

la t ~

rR4" R E
~ South Wind

North
j r'j
~~
I 1J

STEP 18: Bed alcove for

grandmother
Small Kitchen Large Kitchen
If the family has asked for a
bed alcove (choice 7) down-
stairs , it is placed in the
kitchen patio, next to the
family room. Draw according
to table .

-46-
STEP 19: Position of shop

If the family has asked for a shop (choice 11) the house either
fronts on the market, or has a corner lot. The e~act arrange-
ment depends on the relative positions of the corner, the sala,
the kitchen. Draw according to table.

Medium or
Tiny Sala Large Sala

Shop facing market

Medium or Medium or
Tiny Sala Tiny Sala Large Sala Large Sala
East Corner west Corner East Corner West Corner

Shop at front of house

~~~
. I t. :='J: C -
6.· ~.
I w··.
• •

Small Kitchen Small Kitchen Large lli tchen Large Kitchen

East Corner Weist Corner East Corner West Corner

Shop at back of house

STEP 20: Back door and rental

space Open on
Back.
Ope~
west
on Open on
East

If the back or the side of the

~~~
house is open to a pedestrian

~q ~~ ~u
way, the back door is placed
according to the site and the Small Kitchen

position of the kitchen. If

the family has asked for rental

~····-~ ~ ~
space (choice 10) this is built
in the position marked R; other-
wise this position is left
unroof.ed. Draw according to
table.

-47-
 /
.

COSTS

The generic house, shown on pages 8-9, will cost 119,000

soles ($2800) as of summer 1969. The smallest house, shown

opposite, will cost 79,000 soles ($1800), and the largest,

with all possible extras, also shown opposite, will cost

163,000 soles ($3800). These costs are within 1000 soles of

the targets set by the United Nations. They give an average

of 1130 soles per square meter of interior space (not includ-

ing verandas or overhangs). This is 25% less than current low

cost construction in Lima.

Our major cost savings have come from the following

sources: The foundation slab, without footings, costs 100

2 2
soles/m , compared with the usual price of 200 soles/m for

slab and footings. The mortarless concrete block walls rein-

2
forced with sulphur, costs 120 soles/m , compared with the
2
usual price of 140 soles/m for a mortared block or brick wall.

The long side walls are two leaf party walls, thus halving the

usual cost of individually owned walls. The bamboo-urethane

2
floors and roofs cost 200 soles/m , compared with the usual
2
cost of 340 soles/m for reinforced concrete slabs. The fin-

ish of the mortarless block wall and the finish of the bamboo

ceilings make p·lastering unnecessary, and save the usual cost

of 50-60 soles/m
2 for plaster. The dacron sailcloth cover on

-48-
LARGE HOUSE

OOOClD
0 0
a
D

Second Floor

10 3.

First Floor

SMALL HOUSE

l Entrance
2 Sa la (Par l or)
3 Family Rocrr.
4 Fami l y Room Alcove
5 Main Patio
6 Veranda
7 Kitchen
8 Laundry
9 Kitchen Patio Second Floor
10 Stora ge Pat i o
11 Master Bedrcom
12 Bed Alcoves
13 Mirador
14 Clothes Drying Closet
15 Shcwer
16 Toi l et
17 Future Bed Alcove

2 METERS

First Floor
-49-
 2
the patio, costs 250 soles/m and saves the cost of windows
2
throughout the house, at a usual cost of 500-600 soles/m .

We have eliminated several doors at a cost of 550 soles per

door. The ABS accumulator and use of the cavity wall as a

vent, saves the cost of several meters of waste pipe, vent

pipe, and connections. The fact that our site plan has

1726 houses, as against the 1500 expected, saves 12% .of the

cost of site development.

All costs are for summer 1969. The savings are based

on innovations in the building and site development only,

since land and financing costs were fixed.

-50-
THE PATTERNS

Introduction 53

Community Patterns
Cells 55
Road System 63
Pedestrian Network 71
Community Spine 87
Cell Interior 99

House Patterns

House Shape and Orientation 113

Public Part of House 127
Family Part of House 139
Sleeping Areas 153
Entrance and Facade 167
Patio Section 181
Shop and Rental 191

Construction Patterns 195

INTRODUCTION

We now present the sixty-seven patterns from which we have

built our designs. We do not present these patterns merely to

explain our designs, but because we believe that each of them

expresses a generally valid principle, which can be used over

and over again.

----
-This is the essential point of the patterns:
-·- - -- --· --

-----------
they are re-usable. Since many of them deal specifically with

Peru, we hope that they may be particularly useful to Peruvian

architects and builders.

A pattern defines an arrangement of parts in the environ-

ment, which is needed to solve a recurrent social, psychological,

or technical problem. Each pattern has three very clearly de-

fined sections: context, solution and problem.

The context defines a set of conditions. The problem

-·---------- --·- ------ --

defines a complex of needs which always occurs in the given

context. The solution defines the spatial arrangement of parts

which must be present in the given context in order to solve

the problem.

If the needs in the problem are corr ect, and do occur as

stated in the given context, then this arrangement of parts, or

an equivalent one, must always be included in any design for

-53-
 the given context. ~ny design for this context which does not

~ include the pattern, is failing to so ~ ve a known problem.

This does not mean, of course, that the patterns are

absolute. The rightness or wrongness of a pattern is an em-

pirical matter, and as such is always open to further observa-

tion and e x periment . For this reason, we have tried to state

the observations and evidence behind the patterns as clearly

as possible, so that they can be checked by others , and rejected

when incorrect.

/""''\_.,- The evidence we µse comes from three sources : the

<
-·---
published literature , -our -ob~ervations in Lima , and our labora-
-~

tory ~ests and experiments ~ We spent a month each living with

low income Peruvian families in Pampa de Comas , San Martin de

j,
Porras , La Victoria and Rimac (districts of Lima) to better

understand their way of life . We built and tested each of the

major building components , with supportive testing from pro-

fessional laboratories. Where our observations are hard to

support , we have stated them as conjectures .

We preface each pattern statement with a description of

the way the pattern enters into our designs . This preface al-

ways begins with the words : IN THE PROYECTO EXPERIMENTAL ...

The pattern itself follows , beginning with the words : THE

GENERAL PATTERN IS
.I ',-.,.
'\_ I
L, ,

-54-
 CELLS

SUBCULTURE CELLS

DEGREES OF PUBLICNESS
SUBCULTURE CEL LS

IN THE PROYECTO EXPERIMENTAL, the community is divided into 43

small residential cells, each containing between 25 and 75

houses. The cells are clearly separated from one another. All

houses in a cell face ~nwards, and the outer cells are surrounded

by a narrow road sunken 50 cm below grade , so that these cells

are elevated pedestrian islands .

Families choose the cell they want to be in , according to

its relative "quietnes-s", and _according to the community facil-

ities nearby. As a result, the families in any one cell will

probably share attitudes and interests; we hope that each cell

will develop a unique "character ", different from the others .

THE GENERAL PATTERN IS :

Context :

Any qrban area which contains more than a few hundred

dwellings .

-56-
 Solution:

The area is made up of a

large number of small inward

focussed residential "cells".

The cells are separated as

sharply as possible from one

another , if possible, by open

land, community facilities, or

public land.

Each cell is intended, in the long run, to sustain a dif-

ferent way of life: a different subculture. A subculture is

defined as a group of people (not necessarily friends) who

share certain attitudes, beliefs, habits and needs not shared

by others, and who may require special environments, local

organizations, or services, to support these special needs.

The community facilities which surround any given cell should

reflect the particular interests characteristic of that sub-

culture. All community facilities (including roads, schools,

hospitals, churches, parks, industry, commerce, entertainment)

are placed in the boundaries between cells .

The arguments which define cell size, are not yet fully

clear. At present it seems that no cell should contain more

I. than 1500 people, or less than 50 , with a mean cell population

of about 500.
j,
-. j J (V'.J
, . I
r
l;,' ,"'
• •
I
~ ..J
~ " '"1.· . V.:J-
,,,
,,.,
l,Q ~'- ( __:C>-r"YI ,/) 6_ ~)I 1&_ - c
-57- J
 /
.

Problem:

People need an identifiable unit to belong to. They want

to be able to identify the part of the city where they live,
as distinct from all others. Available evidence suggests that
the areas which people identify with are extremely small - of
the order between 10-0 and 200 meters in diameter. They cannot
identify these areas, unless the areas are well differentiated
from one another: and studies show that areas will not be
strongly differentiated from one another unless they support
identifiably different ways of life. This suggests that any
urban area should be broken into a number of small , "subculture
cells'', each supporting an identifiably different way of life .
See Frank Hendricks, "A Situational Approach to Residential
Environmental Planning: A Research Framework" , unpublished
report to the U.S . Public Health Service , March 1967.

Psychological arguments lead to the same conclusion .

There is strong evidence to suggest that a person cannot de-
velop his own life style fully, unless he does so in an
ambience where others share his life style .

In a homogeneous urban area, differences of life style

tend to vanish, and ego-strength , self-confidence and char-
acter formation deteriorate . This again , suggests that the
 urban area, should, as far as possible, support a large variety
of strongly differentiated life styles, each supported by a
"subculture cell".

Ecological arguments help to fix the suitable cell size,

and the need for radical separation between cells. To develop
their own life style, the families in a cell must be able to
agree on basic decisions about services, community land, etc.
Anthropological evidence shows that a human group cannot main-
tain the face to face relations required to coordinate itself
in this way, if its population is above 1500; many people set
the figure as low as 500. (See for example, Anthony Wallace,
Housing and Social Structure, Philadelphia· Housing Authority,
1952; currently available through University Microfilms, Inc.,
Ann Arbor, Michigan.)

It has also been shown that the group feeling necessary to

support a particular unique life style, is greatly strengthened
when that group is physically separated from all adjacent groups.
This suggests that cells should be inward looking, and wherever
possible separated by community facilities.

The ' full arguments, and empirical evidence for all these
points, are presented in Christopher Alexander, Cells of Sub-
cultures, Center for Environmental Structure, Berkeley,
California, 1968.

.../-,.-:.,.I ... -,

'' I['_

~.... '/ ..
r, / ,,_,
,'

-59-
DEGREES OF PUBLICNESS

IN THE PROYECTO EXPERIMENTAL, 30% of the houses are on heavily

travelled pedestrian paths and close to the activity nuclei

(see page 75), 30% are in very secluded areas on small pedes-

trian paths that few people will go on, and the rest are in

betwe'en.

THE GENERAL PATTERN IS:

Context:

Any local urban area which contains more than a few

hundred houses.

Solution:

Houses in different parts of this area, have markedly

different degrees of exposure to pedestrian circulation and

nearby activities. Some are secluded, far from' activity, on

narrow lanes, with few people going past. Others are on main

pedestrian thoroughfares, within hailing distance of potential

activity, with many strangers going past all the time.

-60-
Problem :

It has been shown that people vary along a fundamental

personality characteristic which may be called "introvert-
extrovert" or "privacy loving-community loving" . Those at the
introvert end of the scale, have a consistent desire to be re-
moved from services, and from passing foot traffic, and enjoy
a very small scale in their immediate surroundings. Those at
the extrovert end of the scale like to be near services, to
have a lively atmosphere outside their houses, and enjoy a
larger scale of public spaces. See, for instance, Francis
Loetterle, Environmental Attitudes and Social Life in Santa
Clara County, dittoed manuscript, County of Santa Clara,
County Planning Department, San Jose, California, 1967; and
Frank Hendricks, "A Situational Approach to Residential Environ-
mental Planning", unpublished report to the U.S. Public Health
Service, March 1967.

While it is of course true in every plan, that some houses

are far from services, and others closer to them, the character
of the houses at different distances does not always support
the psychological ne~ds defined above. To support these needs,
the community layout, close to services, should be open to
activities and crowds and large_ in scale, while the community
layout further from services, should be secluded and private
and small in scale.

-61-
THE ROAD SYSTEM

LOOPED LOCAL ROADS

T-JUNCTIONS

DIRECT VISIBLE PARKING

TINY PARKING LOTS

LOOPED LOCAL ROADS

IN THE PROYECTO EXPERIMENTAL, all access to houses is provided

by one lane , one-way , loop roads . No one of these loops ser-

ves more than 100 houses or 50 parking spaces (at 50% car

ownership) .

THE GENERAL PATTERN IS:

Context:

Any residential area served by local roads .

Solution:

These local roads are narrow one-way "loop" roads serv-

ing a maximum of 50 parking spaces . They need be no more than

one lane wide , the surface should be rough . A loop road is

defined as any road in a road network placed so that no path

along other roads in the network can be shortened by travel

along the "loop".

Problem:

Through traffic is fast, noisy, and dangerous. At the

same time , cars are important, and cannot be excluded alto-
gether from the areas where people live. To safeguard these

-64-
areas, the roads must be laid out to discourage all through
traffic - hence the loops. The loops themselves must be de-
signed to discourage high volumes or high speeds: this
depends on the total number of houses served by a loop, the
road surface, the road width, and the number of curves and
corners. Our informal observations suggest that a loop is,
and feels, safe so long as it serves less than 50 cars. At
this level, there may be a car every two minutes at rush
hour, and far fewer during the rest of the day. The number
of houses served will vary, according to the average number
of cars per house. At l~ cars per house, such a loop serves
30 houses; at 1 car per house 50 houses; at ~ car per house,
100 houses.

-65-
T-JUNCTIONS

IN THE PROYECTO EXPERIMENTAL,

no two roads cross. All roads

meet in T-junctions.

THE GENERAL PATTERN IS:

Context:

Any point in a road network where two roads meet, with-

out a traffic light.

Solution:

The roads meet in a T-junction, whose angle is as near

90° as possible.

Problem:

Studies show that traffic accidents are far more frequent

where two roads cross, than at T-junctions. This follows from
the geometry. Where two two-way roads cross, there are 16
collision points, compared with 3 for a T-junction. (John
Callendar, Time Saver Standards, Fourth Edition, New York,
1966, p. 1230.)

Further evidence shows that

the T-junction is safest if it
is a right-angled junction.
When the angle deviates from
the right angle, vision is less
good, and there is confusion
about ·right of way . Accidents
increase. (The SCAFTS Guide-
lines, Swedish National Board
of Urban Planning, 1968, ~- 11.)

-66-
 DIRECT VISIBLE PARKING

IN THE PROYECTO EXPERIMENTAL, every house is within 50 meters

of a parking lot, and in most cases the parking lot is visible

from the front of the house .

1'1
~I 0Ii H I ! Ii ~ 11
l • t·
\

'
,, 1f
i; ! \
r'.

i ' L

. THE GENERAL PATTERN IS :

Context :

Parking for a housing area where car ownership is unpre-

dictable , and less than one car per house.

-67-
Solution:

There is a parking lot directly visible from every house,

and less than 50 meters away from every house. If the house

has two doors, then the parking lot is closer to the front

door than to the back door.

Problem:

Among low income people, who have few cars, the car is a
prized possession. Often it may be the most valuable thing
its owner has.

A housing project in which car parks are far away from the
house, and hidden from the house, is entirely unacceptable to
such a person. He wants to be able to watch his car, to make
sure no one steals it, and he wants it to be associated with
his house because he is proud of it. The car must be close to
the house, visible from it, and visibly associated with it.

It is also essential that the car be associated with the

front door of the house, not the back door. When people have
cars, they always come in and out of the house by the door

-68-
which is closest to the car.
(See, for instance, Vere Hole ,
et al, "Studies of 800 Houses
in Conventional and Radburn
Layouts", building Research
Station , Garston , England ,
1966 . ) If t he relationship
between car and house is such
that this leads them to use
the back door of the house,
this turns the internal or-
ganization of the house in-
side out . The classic example
of this mistake , is in Radburn
itself , where all houses have
access to the parking lot
through the back door. The
back door becomes the front
door . Everyone, including
visitors , comes into the hous~
through the back door .

In a country like Peru ,

where t he distinction between
the front door and the back
door is so crucial socially ,
this kind of mistake would be
disastrous . (See I ntimacy
Gradient and Puerta Falsa
Patterns . ) The parking lot
must be nearer to the front
door than the back door . Since people will have to carry heavy
parcels , 50 meters is a widely accepted upper limit for this
distance . (E . g ., Kevin Lynch , Site Planning , M. I . T . Press ,
1962 , p . 181.)

-69-
TINY PARKING LOTS

t
EVERY CELL IN THE PROYECTO EXPERIMENTAL, is surrounded by a

number of small parking lots, each one designed for 3-7 cars.

T.HE GENERAL PATTERN IS :

Context:

Any open part of a community containing parking.

Solution:

The parking is broken down into separate , tiny parking

lots , no one parking lot holding more than 6 cars .

Problem :

Large parking lots create an impersonal , institutional

atmosphere , dominated by cars ; they separate people from the
pleasure and convenience of being near their cars ; and , if
they are large enough to contain unpredictable traffic, they
are dangerous for children , since children inevitably play
in parking lots .

It is hard to pin down the exact size at which parking

lots become too big . Our informal observations suggest that
parking lots for four cars are still essentially pedestrian
and human in character; that lots for six cars are acceptable ;
but that any area near a parking lot which holds eight cars ,
is already clearly identifiable as "car dominated territory''.

This may~ be connected with the well-known perceptual facts

about the number seven . A collection of less than 5-7 objects
can be grasped as one thing, and the objects in it can be'
grasped as individuals. A collection of more than 5-7 things,
is perceived as "many things". (See G. Miller, "The Magical
Number Seven, Plus or Minus Two: Some Limits on Our Capacity
for Processing Information", in D. Beardslee and M~ Wertheimer
(eds.) Readings in Perception, New York, 1958, esp. p . 103.)

It may be true that the impression .of a ."sea of cars"

first comes into being with about seven cars~

-70-
PEDESTRIAN NETWORK

PASEO

A~TIVITY NUCLEI

CAR-PEDESTRIAN SYMBIOSIS

PEDESTRIAN 50 CM ABOVE CAR

KNUCKLE AT ROAD CROSSING

PASEO

IN THE PROYECTO EXPERIMENTAL, there is a curving "paseo" down

the middle of the site, coming out to the artery at the ends.

This paseo is a sequence of looped paths, connected by a se-

quence of small squares. All community facilities open onto

these small squares. Between these squares, the paths pass

between terraces of houses. Two-thirds of the houses in the

project, are within 100 meters, and all are within 180 meters,

of the paseo.

THE GENERAL PATTERN IS:

Context:

Any residential area, in a Latin country.

Solution:
5
·. The area contains a "paseo" . We defl.n e a paseo as a

system of looped pedestrian paths, connecting a system of

activity nuclei. Those activity nuclei at the ends of the

paseo contain facilities which can function as destinations

for an evening walk: ice cream parlors , coke shops, church,

-72-
 volley ball courts, public gardens, movie house, bars, swim-

ming pools . No point on the paseo is more than 50 meters from

an activity nucleus, and no house is more than 100 meters from

the paseo. Where the paseo is not fronting on activity nuclei,

it always fronts on houses ~ it never fronts on blank walls, or

the side walls of houses.

Problem:

In Latin America, Spain, Sicily, Italy, the habit of the

evening "paseo" is a very common one . On many evenings, and
in the daytime on Sundays and holidays, people go to some cen-
tral place where they walk up and down, looking out for friends,
enjoying the talking as they walk, showing themselves.

l
 Since people come to see people, and to be seen, this
place must have a high density of pedestrians using it. It
must therefore be associated with the places most likely to
attract people: the activity nuclei.

Further, even though the real reason for the paseo has to
do with seeing people, and being seen by them, people in gen-
eral find it easier to take such a walk if they have a "des-
tination" to go to. This destination may be a real destination
(like a coke-shop or cafe) - or it may be partly imaginary,
like "lets walk round the block" - which again provides a per-
son with a specific goal.

It is also important that people do not have to walk too

far between the public activities. Informal observation sug-
gests that any point which is more than 50 meters from activity
becomes unsavory and unused. By the same token, no house
should be too far from the paseo: if people have to go through
deserted streets at night, to get to the paseo, the chances are
that they won't go there. 100 meters seems to be a reasonable
upper limit.

For a general description of the paseo, see Edward B. Allen,

"The Passegiata", Landscape, Winter 1969, pp. 29-32.

-74-
ACTIVITY NUCLEI

IN THE PROYECTO EXPERIMENTAL, all community facilities open

onto one of eight small squares , and all pedestrian paths in

the community, lead towards these eight small squares .

Each of the activity generators is unique, according to

the facilities which surround it . The market square is sur-

rounded by small shops , has a supermarket at one end , the

artery crossing at the other , and contains market stalls . The

evening center is surrounded by cinema , dance hall , and bars ,

c

and contains clustered lights for night time activity, and

sheltered tables round the edge . The open spaces between kin-

dergartens and walled eucalyptus groves, contain shallow tiled

pools, where toddlers can splash and play, with seats around

them for mothers. The open space between church and clinic

contains flowers , grass and bushes, cared for by the church.

The open spaces in front of primary schools, and sports centers,

contain a stepped depression , large enough for football in the

middle, the steps deep enough to form seats for people who

want to watch the games.

-75-
THE GENERAL PATTERN IS:

Context:

Any community large enough to support conununity facilities.

Solution:

The community facilities are clustered round a small number

of very small open spaces which we call activity nuclei. The

facilities in any one nucleus are clustered in such a way that

they cooperate functionally. (See problem statement . ) All

paths in the conununity pass through these activity nuclei.

Problem:

One of the greatest problems with new communities, is the

fact that the available public life in them is spread so thin ,
that it has no impact on the community , and is not in any real
sense "available" to the members of the community . Yet studies
of pedestrian behavior make it clear that people seek out con-
centrations of other people , whenever they are available , (e . g .
Jan Gehl , "Mennesker til Fods (Pedestrians)", Arkitekten , No .
20 , 1968 . )

To create these concentrations of people in a community ,

facilities must be g r ouped densely round very small public
open spaces which can function as nuclei - and all pedestrian
movement in the community channelled to pass through t h ese
nuclei . These nuclei need two properties .

First , the facilities grouped around any one activity

nucleus, must be carefully chosen for their symbiotic rela-
tionships. It is definitely not enough merely to group com-
munal functions in so called community centers . For example ,
church, cinema, kindergarten , and police station are all com~
munity facilities - but they do not support one another
mutually. Different people go to them , at different times ,
with different things in mind. There is no point in grouping
them together. To create intensity of action, the faci l i t ies
which are placed together round any one nucleus , must f u nction
in a cooperative manner , and must attract the same kinds of
people, at the same times of day .

For example: When evening entertainm~nts are grouped to-

gether, the people who are having a night out can use any one

-76-
of them, and the total concentration of action increases.
When kindergartens and small parks and gardens are grouped to-
gether, mothers and young children may use either, so their
total attraction is increased. When schools and swimming
pools and football space are grouped together, they form
natural centers for school children.

Second, the open places

which form the nuclei should
be very small indeed . Our
observations suggest that
15 x 20 meters is the ideal
size ; if the space has to
be larger , it should be
long and narrow , with its
s h ort d imension no more
than 20 meters .

-77-
 Our observations in Lima showed, again and again, that
places which are larger look and feel deserted, and discourage

activity. The reasons for this recurrent observation are

obscure, but the following facts may have something to do with
it. A person's face is just recognisable at about 20 meters,
and, under typical urban outdoor noise conditions, a loud
voice can just be heard at 20 meters.

This may mean that people feel tied together in spaces

whose diameter is less than 20 meters, and lose this feeling
in larger spaces: perhaps a major factor in the development
of activity.

-78-
CAR -PEDESTRIAN SYMBIOSIS

IN THE PROYECTO EXPERIMENTAL, the car roads form loops, and

the pedestrian paths form a diagonal network which crosses

these loops at right angles. Where they cross, there are park-

ing lots, cell gateways, and space for pedestrian activity.

The two systems form a double gradient: car densities dominate

towards the outside of the site, pedestrian densities dominate

towards the inside of the site , with a smooth gradient between

the two.

THE GENERAL PATTERN IS:

Context:

Any area which contains pedestrian paths and local car roads.

Solution:

The system of pedestrian paths and

the system of roads are two entirely dis-

tinct orthogonal systems. They cross

frequently; so that no point on either

system is more than about 50 meters from

-79-
a crossing. Every time they cross, both paths and roads

swell out, making room for pedestrian activity and for park-

ing and standing.

Problem:

It is common planning practice to separate pedestrians

and cars. This makes pedestrian areas more human, and safer.
However, this practice fails to take account of the fact that
cars and pedestrians also need each other: and that, in fact,
a great deal of urban life occurs at precisely the point where
these two systems meet . Many of the greatest places in cities,
Piccadilly Circus, Times Square, the Champs Elysees, are alive
because they are at places where pedestrians and vehicles meet .
New towns like Cumbernauld, where there is total separation be-
tween the two, seldom have the same sort of liveliness.

The same thing is true at the local residential scale. A

great deal of everyday social life happens where cars and pe-
destrians meet. In many low income areas, for example, the car
is used as an extension of the house . Men, especially , often
sit in parked cars , near their houses, drinking beer and talk-
ing. (Clare Cooper , "Some Social Implications of House and
Site Plan Design at Easter Hill Village: A Case Study ", In-
stitute of Urban and Regional Development , Center for Planning
and Development Research , University of California, Berkeley ,
California , 1966, pp . 39 ff . )

Many studies show that conversation and discussion grow

naturally out of the communal car lots where men mee t when
they take care of their cars . Vendors always set themselves
up where cars and pedestrians meet ; they need all the traff ic
they can get . Children always play in parking lots - perhaps
because they sense that this is the main point of arrival and
departure ; perhaps because they enjoy the cars .
 In Peru, there is a new version of the paseo: the "auto-
paseo" - several friends hop into a car, and drive around,
visiting their friends, often not even getting out of their
cars, but talking from house to car, and back.

None of these things can happen in a plan where car roads

and pedestrian paths are separated, unless the two meet fre-
quently, and the places where they meet are treated as minor
centers of activity.

-81-
 PEDESTRIAN 50 CM ABOVE CAR

IN THE PROYECTO EXPERIMENTAL, all local roads

are 50 cm below grade. The pedestrian pre-

cinct defined by each cell, is an island,

floating 50 cm above the road which sur-

rounds it. Where a pedestrian path crosses

a local road, there are three steps down.

THE GENERAL PATTERN IS:

Context:

Any area which contains roads with traffic densities of

more than a few vehicles per day.

Solution:

These roads are sunk 50 centimeters below all pedestrian

paths.

0
Problem:

In the modern city, the

car is king: the pedestJ:ian is
made to feel. small. This can-
not be solved by keeping pe-
destrians separate from cars;
it is · in their nature , that
they .have to meet. But where
they meet, the car must be
wput down", symbolically, and
the pedestrian world g ,i ven '··
more importance. This i§ mdst
easily achieved if. the car is physically below the pedestrian.
our experiments suggest that the ·e ffect first makes itself
felt ,, when the. car is about 50 ·Centimeters bel-0w th~ pe~es­
trian paths. There are two possible reasons for t.h1s figure.

Most people's eye level is

between 1.30 and 1.60 meters.
A typical car has an overall
height of 1. 40 meters,. Al- .70
though tall people cg.n see ___
..........

over the cars, even fo:r: them

the car:s fil! the landscape,
since a standing person's nor- 4 •.oa
mal lh1e of sight i ·s 10 deg·rees
below the horizontal. (Henry
Dreyfus, The Measure. of Man,
~e~York, 1959, sheet p;) Xo
get the· top of a ~ar that is
four meters away, completely be.low the line of sight, it
wou1'a have t o be ·standing between 50 and 80 centimeters
below the pedestrian.

It may also be that the ·o ar overwhelms the pedestrian

because of a constant, unspoken possibility that a runaway
car might at any moment mount the curb and run him down.
~ car can climb an ordinary 15 centimeter curb, easily. For
the pedestrian to feel c-ertain that a car could not climb
the curb, the curb height would have to be greater than the
radius of a car t:i.re (30-38 cm): thus at least 40 cm, pre- '
ferably 50 cm.

. -B3-
KNUCKLE AT ROAD CROSSING

IN THE PROYECTO EXPERIMENTAL, where the paseo crosses the

artery, the pedestrian path crosses 50 cm above the level of

the roadway, and is marked by a canopy; the artery narrows to

through lanes only, and slopes up to the crossing ; on each

side, the pedestrian path opens out to provide space for com-

mercial activity , with parking bays for loading and unloading .

l I •
I '

\ I

'
•
I

I
t

THE GENERAL PATTERN IS:

Context:

Any point where a pedestrian path crosses a major road .

Solution:

There is a "knuckle" at the crossing, defined by the

f o llowing features:

-84-
1. The road narrows to the width of the through lanes only

and the lanes themselves narrow to 2.40 meters.

2. If the road is more than 3 lanes wide, the roadway is

split in two and the island in the middle is at least

10 meters wide.

3. The pedestrian path continues through the crossing 50 cm

above the roadway.

4. The road slopes up towards the crossing (slope 1 in 6 max . ) .

5. The pedestrian path is marked by a canopy, or shelter, ex-

cept where it crosses the roadway.

6. Immediately next to the point of crossing, there are bays

opening off the roadway , providing standing spaces for

busses, cars, delivery trucks.

There are steps leading to these

standing bays.

7. On one; or both sides of the

road, the pedestrian path swells

out to make space for shops, and

vendors.

Problem:

The way a path crosses a road, depends on traf.:f,±c- density .

If the traffic density is very low, it is safe to let people
cross where they want to. If the traffic density is high, the
crossing has to be specially defined.

The main problem for pedestrians trying to cross a heavily

travelled road, is the width of the roadway. Buchanan has
shown that the average waiting time, and the percentage of
pedestrians who are forced to wait, for various levels of

-85-
traffic flow, is greatly" affe.cted by the width of the road.
(Co·lin Buchanan, e ·t al, .T raffic: in Towns, HMSO, London .,. 1963,
pp. 203-213.) The. road and the lanes themselves, should
therefore narrow as far as poss.i b.le at the cro.ssing. Any
crossing more ,t han three lanes wide, should be split in two.

The. fa.ct that pedestrian aS'cendancy over the car Will be

ine.reasea if they are 50 centimeters higher, has already been
discussed in "Pedestrian 50 cm Above, Car 0 • The same principle
applies, even more. powerfully, where pedestrians have to cross 4
a road. 'l'he pedestrians who cross, must be. extremely visible
from the roaa. Cars should also be forced to slow down when
they approach such a crossing . If the pedestrian way crosses
above the roadway, and the roadway slopes up to it, this sat-
isfies both reguirement,s . Small bumps in the road ., to slow
·cars down. at pedestrian crossings, are commo·n ln Peru. The
slope may not 'b e too steep, since this· wouid be dangerous. to
traffic, and make the pedestrians invisible as the car nose
rides up. 'A slope of; 1. in 6, or .less, is safe from both points
of view:.
To make the crossing even easier to see from a distance,
and to give weight to the pedestrian's "right to be there" , the-
pedestrian path should be marked by a canopy af: the edge o f •
the road.

Finally,, bus stops, vendors, shops, grow naturally around

a place where a pedestrian path cr'oSses a road. There must
be e·nough room. for these activities on both sides of the cross-
ing, and for the parl<:ing, standing, lo~ding and unloaaing that
go with them.

-86-
COMMUNITY SPINE

CENTRAL MARKBT

EVENING CENTERS

WA.L K-THROUGH SCHOOLS

VI SIBLE KI:NbERGAR'!rEN S

SOCIAL GARDEN'S
CENTRAL MARKET

IN THE PROYECTO EXPERIMENTAL , the market and supermarket , are

placed at that point where the paseo meets the northbound

artery which crosses the site . The market its elf is. in a long

narrow open space, lined on both sides by small shops , with

opportunities for a double row of open stalls down the middle .

THE GENERAL PATTERN IS:

Context:

Any - rcrw-·income community containing an open market.

Solution:

The market is at the center of gravity of the community,

next to a major traffic artery, in a pedestrian area running

back from the artery.

Problem :

In areas where there are few cars and few refrigerators,

women buy the food fresh, every day, and carry the food home
on foot . It is therefore essential that the market be located
in the most central position in the community, a position which
minimizes total walking distance for the entire community .

On the other hand, many community market centers benefit

from direct access to urban arterials. The central market must
therefore be located along that section of an artery which is
most central to the community which it serves.

The market area should be at a point of high pedestrian

flow, and shops should be oriented alongside the pedestrian
flow so as to maximize exposure to potential consumers. The
market stalls must also be in a pedestrian area - since most
shopping at a market is comparison shopping, and will continue
to be so, even when supermarkets take over .

Finally, market shops need easy access of vehicles for

deliveries. However, since deliveries are infrequent, trucks
and vans can move through pedestrian areas when delivering .

-89-
EVENING CENTERS

IN THE PROYECTO EXPERIMENTAL, the cinema, dance hall for Gran

Baile Social, bars, restaurants, police station, and main bus

stop are clustered round an open space, which is filled with

lights, where the paseo meets the northbound artery that

crosses the site.

THE GENERAL PATTERN IS:

Context:

Any urban area which contains community facilities that

are open at night (including, for instance, bars, restaurants,

cinemas, police stations, bus stops, all night gas stations).

Solution:

These evening facilities are aggregated to form clusters,

no one cluster containing less than half a dozen facilities.

-90-
Problem :

1. People feel unsafe and uncomfortable at night when they

are not grouped in large numbers . When large numbers of
people are in sight people are comfortable .

2. If evening activities such as police station , cinema ,

bars , cafes , ice cream parlors , gas stations and the like
are scattered throughout the community , each one by it-
self cannot generate enough attraction for community mem-
bers . If activities are grouped , the patrons of one ac-
tivity are also potential patrons of the others . The
total attraction of customers of each activity is thus
increased . Furthermore , the nucleation of evening ac-
tivities , creates larger concentrations of people , and
encourages new evening activities to open there .

3. Many people do not go out at night because they feel they

have no place to go , and they do not feel like going out
to a specific establishment . But they feel like going
out . An evening center , particularly when it is full of
light , functions as a focus for such people , and by attrac-
ting them, again increases its potential for liveliness and
action .

The subject of evening centers has been discussed at length

by Shlomo Angel, "Discouraging Crime Through City Planning" ,
Working Paper No. 67 , Center for Planning and Development Re-
search, University of California, Berkeley, 1968.

-91-
WALK - THROUGH SCHOOLS

IN THE PROYECTO EXPERIMENTAL , secondary schools are placed at

the ends of the paseo , with main pedestrian arcades runni n g

through them. These two schools will be laid out so that all

facilities which might possibly become used for adult educa-

tion , or for invited gatherings in the school , li k e school

plays, will be located along these main arcades .

/

THE GENERAL PATTERN IS:

Context:

Any school in a community.

Solution :

The schools are ·laid out so that public pedestrian paths

run through them, and those parts of the schools which may

-92-
involve the wider public (like playgrounds, auditoria, gardens,

workshops) are placed to open directly off these paths.

Problem:

There is a general tendency, in modern society, to dis-

sociate the formal education of children from the other
processes of society . Children spend the day in a school,
which is separate from most ordinary social processes, and then
spend the night at home . This contrasts very badly with the
situation in a non-urban society , where a child is exposed to
all social institutions - work, play, ritual, government - all
the time. It is not surprising that many young people in urban
society no longer understand their relation to society .

In order to combat this dissociation, it is essential that

schools be ''opened" in the most literal sense, to the rest of
the community . Of course , active participation of community
persons, in the daily life of the school, cannot be quaranteed
by physical layout . But it will not happen , unless the physical
layout helps to encourage it . In order to destroy the self-
contained , institutional character of the schools , they should
overlap , as far as possible, with public thoroughfares which
encourage publi c use .

-93-
VI S IBLE KINDERGARTENS

IN THE PROYECTO EXPERIMENTAL , each kindergarten is a walled

precinct , containing sunken play areas . Public paths cut

through this precinct alongside the sunken play areas , and

are separated from these areas by a low seat-height wall .

I
'
I 1 '
- .. '' I

I
- t I
'- I I , ,,,.
.J.
J/ ''I .
-
! i- I , f, p ; /
.i - J 11 . ) . IJ I
' .J 1 ! v r' I

;! I i

\
/

-94-
THE GENERAL PATTERN IS:

Context:

A residential community containing kindergartens within

walking distance of the houses they serve.

Solution:
I .
Public paths pass through the kindergartens, and the open

areas inside the kindergarten are sunken, adjacent to the paths,

and separated from them by a low seat-height wall.

Problem:

There is a general tendency, in modern society, to dis-

sociate the formal education of children from the other
processes of society. This dissociation leads, in the end,
to a situation where young people no longer understand their
relation to society. (See Walk Through Schools.)

The first step of this process is the "trauma" which the

young child experiences when first taken off to formal "school" -
i.e., kindergarten. Silverstein has indicated that the child's
sense of this first school being ''separate" from society, can
be reduced if the play areas of the kindergarten are open to
all passing adults, and to all passing children - not only
those who happen to be enrolled. (Murray Silverstein, "The
Child's Urban Environment", Proceedings of the Seventy-First
National Convention of the Congress of Parents and Teachers,
Chicago, Illinois, 1967, pages 39-45.)

To keep the young children safe, and to make it possible

to give them this great freedom without losing track of them
altogether, the play areas may be sunk slightly, and sur-
rounded by a low wall. If the wall is at seat height, it will
encourage people to sit on it - giving them a place from where
to watch the children playing, and the children " a~ chance to
talk to passers-by.
WALLED GARDEN S

IN THE PROYECTO EXPERIMENTAL , the re are three small walled

gardens cont a ining g ra ss , seats , paved areas and eucalyp tus

trees . Each one opens onto o n e of the activity nuclei along

the paseo . The eucalyptus will be a fast growin g variety ,

suitable fo r Peru , and should reach a h eight of 25 meters in

f ive years . The gardens will be irrigated by the irrigation

water that comes in from the northeast corner of the s ite .

THE GENERAL PATTERN IS :

Context :

Any small park or public garden in an urban area.

Solution :

It is walled, or partly enclosed, yet close to major cen-

ters of pedestrian activity.

Problem:

People need contact with trees and plants and water. Their
symbolic character is not replaceable. In some way, which is
hard to express, people are able to be more whole in the pres-
ence of nature, are able to go deeper into themselves, and are
somehow able to draw sustaining energy from the life of plants
and trees and water.

The small parks and gardens in a city try to solve this

problem; but they are usually so close to traffic, noise , and
buildings, that the impact of nature is entirely lost . To
be truly useful, in the deepest psychological sense , they must
allow the people in them to be in touch with nature - and must
be shielded from the sight and sound of passing traffic.

In those few cases where there are small walled gardens in

a city, open to the public - Alhambra, Morocco, Copenhagen Royal

-96-
 J
Library Garden - these gardens almost always become famous .
People understand , and value the peace which they create .

This is a particularly crucial problem in desert areas like

Lima . In the desert , trees and plants are infinitely precious .
Gardens are almost like oases - people flock to sit and talk
and lie in them , wherever they e x ist . In such desert areas it
is doubly important to keep at least partial separation between
the garden and its surroundings , so that the garden can be
fully felt , and the oasis character is n' t lost .

At the same time , just because they are so precious , such

gardens need to be close to major centers of pedestrian activity,
so that people can use them and enjoy them often .

I ~

I
 L

CELL INTERIOR

CELL GATEWAY

MULTIPURPOSE OUTDOOR ROOMS

SHOPS ON CORNERS

CENTRIPETAL PEDESTRIAN PATHS

STREET FOOTBALL

FLOWERS ON THE STREET

CELL GATEWAY

IN THE PROYECTO EXPERIMENTAL , every cell e ntrance is defined

by the f act tha t the p e de s trian path g o e s down three steps , to

cro s s a road, then up three s teps again , and the n pas ses under

a woven b amb oo canop y . In many cases the p a th al s o nar r ows at

the e n t rance .

!
I
.
;?
;
I
.1

--; ·-·----

THE GENERAL PATTERN IS:

Context:

Any point where a pedestrian path enters a residential

cell, as defined by the subculture cells pattern.

-100-
Solution:

The path n a rrows down ,

changes level , p a sses under

,,
cover or passes through an

e n clos ur e .

Problem:

The a r g ument s in t he s ubcu l ture c el l pattern exp l ain t h at

the h o u ses n e ed to b e g rouped in cells which can d evelo p a
stro ng ly ind ividua l ide ntity and atmosphere . This wi ll only
h app e n i f the p eop l e wh o live there , contri bute to t h is atmos-
phere ; a nd this in tur n , wi ll o nly happen if they experience
the di s t i nctne ss of thei r own ce ll , dai l y .

Peop le d o not e x peri e nc e t h e d i s t inctness of an area , un-

l e ss many d iffer e nt qua l itie s c h ange simultaneously , as they
e n ter that area . Th is requires a zone of transition , a t every
point wher e a path enters the cell , with as many environmental
qualities as possib le , changing during passage through this
transition zone .

If possible the sound of footfall, the view, the feel of

the surface , the level, the relative amount of enclosure, the
intensity of light should all change together. The transition
will be best marked of all, if it is experienced as a "bridge"
or a "gate".

There is little published evidence for this phenomenon -

though the Japanese, for instance, have made use of it for
centuries. The nearest empirical evidence is Kevin Lynch's
demonstration that "edges" are major components in the images
of cities which people have. (The Image of the City, M.I.T.
Press, 1960, p. 62 ff.)

-101-
MULTI-PURPOSE OUTDOOR ROOM

IN THE PROYECTO EXPERIMENTAL, each cell contains one open

space about 6 x 10 meters, surrounded by a two meter deep,

roofed, arcade, and several smaller, non-continuous sections

of the same arcade in other places. Each one of these arcades

is placed at a node in the system of pedestrian paths. They

are left unfinished, with the understanding that community

residents may build in games like table tennis, bochas or sapo,

sand-pits, seats, water faucets, walls, and small shops or bars,

according to their needs.

I(
1,
I / I !1

, I I

THE GENERAL PATTERN IS:

Context:

Any local part of a residential community.

-102-
Solution:

There is, within view of every house, at least one

"multipurpose outdoor room" with the following characteristics.

It is open to the sky, surrounded or at least partly surrounded

by a continuous roofed arcade, always at least two meters deep,

and, where possible, built up against the walls of existing

buildings.

...

The outdoor rooms are left unfinished, with the under-

standing that they will be finished by people who live near

them, to fill whatever needs seem to be most pressing. They

may contain sand, or water faucets, or play equipment, for

small children. They may contain steps, and seats, where

teenagers can meet. Someone may build a small bar in a house

that opens into the arcade, making the arcade a place to eat

and drink. There may be games for old people, like chess and

checkers.

-103-
Problem:

In existing modern housing

projects, people rarely feel
comfortable lingering outside
their houses . There are few
places where it is "a ll right
to be". Yet at the same time ,
it is clear that almost every-
one wants , at some time or an-
other, to linger in some public
space . Our observations in Peru
show that the men seek corner beer shops , where they spend
hours talking and drinking; teenagers , especially boys , cho os e
special corners too , where they hang around , waiting for their
frie nd s . And of course thes e things are not peculiar to Peru -
something like it happens everywhere . Old people like a special
spot to go to , where they can expect to find others ; small child-
ren need sand lots , mud , plant s , and water to play with in t h e
open; young mot h ers who go to watch their children , ofte n use
the childreds p lay as an opportunity to mee t and talk with oth er
mothers .

Few modern housing projects provide fo r these needs ; it is

very hard to provide for them . On the one hand , indoor com-
munity rooms are too enclosed . When provided , they are rarely
used . People don ' t want to plunge in t o a situation which they
don ' t know; and the degree of involvement created in s uch an
enclosed space, is too intimate to allow a casual passing in-
terest to build up . gradually to full involvement . On the
other hand , vacant land is not enclosed enough . I t tak es years
for anything to happen on vacant land ; it provides t oo little
shelter, and too little " reason to be there " .

What is needed is a framework which is just enough defined

so that people naturally stop there, and tend to stop there ;
and so that curiosity naturally takes people there, and allows
them to stay there . Then, once community groups begin to grav-
itate towards this framework , there is a good chance that they
will themselves create an environment which is appropriate to
their activities. Some possible e xamples of such future de-
velopments are given in the solution statement .

We conjecture that a small open space , between 10 and 20

meters in diameter , and surrounded on all sides by an open
roofed arcade, may just about provide the necessary balance
of "openness" and "closedness". The arcade should be at least
2 meters deep , for the reasons give n in "Two Meter Balcony ".

Even if this conjecture turns out to be correct in theory ,

it will undoubtedly be very hard to implement . Only detailed
experiments, in communities , will show up the finer points
that are needed to make this pattern work in practice .

-104-
SH 0 PS ON CORNERS

IN THE PROYECTO EXPERIMENTAL, all

shops which are not located in the

central market, and all houses whose

owners have expressed a desire to open

a shop in the future, are located on

key corners which command a view up

and down major pedestrian paths.

..
""'

THE GENERAL PATTERN IS:

Context:

An urban area contains scattered shops.

Solution:

The shops are at poi'nts of maximum pedestrian density, on

corners.
Problem:

Survival is a key issue for any small shop. The success

of a shop depends on its location; it should always be placed
at a point where the largest possible number of people are
going past, and where it can be seen from many directions.

It has been shown that the rents which owners of small

retail businesses are willing to pay, vary directly with the
amount of pedestrian traffic passing by, and are uniformly
higher on street corners, than in the middle of the block.
(Brian J. L. Berry, Geography of Market Centers and Retail
Distribution, Prentice Hall, 1967, p. 49.)

-106-
 CENTRIPETAL PEDESTRIAN PATHS

IN THE PROYECTO EXPERIMENTAL, the pedestrian ways outside

houses are convex in plan, have seats and galleries round

the edges, and are roofed over by esteras (woven bamboo

matting) at the cell gateways and in the middle of the cell.

THE GENERAL PATTERN IS:

Context:

Any pedestrian sheet which houses open off.

Solution:

The street is subtly convex in plan, has seats and

galleries round the edges, and is occasionally roofed, per-

haps by a "virtual" roof, made only of beams or trellis-work .

Problem:

For centuries, the street provided city dwellers with

usable public space, right outside their houses. Now, in
a number of subtle ways, the modern city has made streets
which are for "going through", not for "staying in" . This
is reinforced by new regulations which make it a crime to
loiter, by the greater attractions inside the house itself,
and by streets which are so unattractive to stay in, that
they almost force people into their houses .

All this contributes to the fact that people in cities

feel isolated , insecure, detached from society. Two recent
studies have shown that menta l illness, and acute feelings
of isolation , are more common among people who cannot reach
the street from their dwellings , than among those who can .
(D . M. Fanning , "Families in Flats", British Medical Journal ,
18 November 1967 ; and Joan Ash, "Families Living at High
Density" , unpublished mimeographed report, Sociological Re-
search Section , Ministry of Housing and Local Government,
1965.) The fact that the street drives people away from it ,
must surely have the same effect.

-107-
 From an environmental standpoint, the essence of the problem
is this: Streets are "centrifugal" not "centripetal" : they
drive people out, instead of attracting them in. In order to
combat this effect, the pedestrian world outside houses must
be made into the kind of place where you "stay", rather than
the kind of place you "move through". It must, in short, be
made like a kind of outside public room, with a greater sense
of enclosure than a street. It should be partly enclosed, by
convexity, and by partial or virtual roofing; houses should
open towards it, not turn their backs on it; and seats should
be provided round the edges.

-108-
STREET FOOTBALL

IN THE PROYECTO EXPERIMENTAL, cells contain pedestrian places

for football, 8 x 20 meters , protected from the west sun by

houses which run alongside .

THE GENERAL PATTERN IS :

Context :

Any housing area in Peru.

Solution:

Within hailing distance of every house , there is an open

pedestrian space at l east 8 x 20 meters, shaded from the west

sun , with seats , steps or low walls around it .

..
t I ,..._
I

I - - I

-109-
Problem:

Most young people in Peru play football, fubito, or

volleyball in the streets. Any one of these games needs a
space 8 to 10 meters wide and 20 to 30 meters long . In the
summer, the games are often played in the afternoon , and need
protection from the west sun . Since many of these games draw
spectators, they are more enjoyable if surrounded by seats or
steps or walls which people can sit on .

-110-
FLOWERS ON THE STREET

IN THE PROYECTO EXPERIMENTAL, the pedestrian paths outside

houses are paved with large, removable earth-cement paving

blocks laid over unfinished earth. It is understood that any

homeowner who wants to plant flowers or trees or grass outside

his house, may remove these paving stones to do so.

' I
_J.--- --
,_-

l - ... .

THE GENERAL PATTERN IS:

Context:

Any pedestrian path outside houses~

-111-
Solution:

The paths are paved with removable paving stones and it

is understood that the people who live in any given house may

take up the stones outside their house, to plant flowers , trees

and grass.

Problem:

The overall effect of a community, especially in a desert

area like Lima, is largely determined by the planting . An
area with well kept plants is beautiful; the areas without
them seem arid by comparison.

However, the only plants in a community which get looked

after, are those which individual people plant, and care for -
public planting usually flops - no one takes the responsibility
to look after it, and there is no money for gardeners.

We may thus establish a principle: The planting in a com-

munity, should be in the form of small gardens, clearly associ-
ated with the front of specific individual houses, and planted
by the individuals in these houses.

However, only a few of the people in any given community

really enjoy gardening. If all the houses are provided with
front gardens, three quarters
of them will be left unkept -
and will very likely end up as
dust and weeds. There must be
a way of giving gardens to all
those people who will look
after them, and to no one else .
This is easily done . If the
walkways are all paved , with
the understanding that anyone
who wants to can take up the
paving stones to plant things
there , the only people who will
bother to do so , are the people
who really want to have a gar-
den there .

-112 -
 HOUSE SHAPE

AND ORIENTATION

LONG THIN HOUSE

PERIMETER WALL

CROSS VENTILATED HOUSE

LIGHT ON TWO SIDES OF EVERY ROOM

PATIOS WHICH LIVE

TAPESTRY OF LIGHT AND DARK

LONG THIN HOUSE

IN THE PROYECTO EXPERIMENTAL,

all houses are long and thin.

They are always 5.20 meters

wide; their length varies from

13 meters to 27 meters.

THE GENERAL PATTERN IS:

Context:

A small house which contains several people.

Solution:

The interior of the house is so shaped that the mean

distance between rooms is as high as possible. This means

that the inside of the house is, effectively, long and thin,

but the outside of the house may take many physical forms:

it may be S-shaped, U-shaped, tall many storeyed, or it may

itself be long and thin.

Problem:

For a large family living in a small house, the overwhelm-

ing problem is one of overcrowding. They feel cramped. Every-
thing seems to be too near everything else. Privacy, for
individuals, or for small groups, is almost impossible.

-114-
 There is widespread evidence to show that overcrowding in
small dwellings causes psychological and social damage.
(For example, William C. Loring, "Housing Characteristics and
Social Disorganization", Social Problems, January 1956; Chombard
de Lauwe, Famille et Habitation, Editions du Centre National de
la Recherche Scientifique, Paris, 1959; B . Lander , Towards an
Understanding of Juvenile Deliquency, New York, 1954.) 11

The feeling of overcrowding is largely created by the mean I

point-to-point distances inside the house. In a small house

these distances are small - as a result it is not possible to
walk far inside the house, nor to get away from annoying dis-
turbances; and it is hard to get away from noise sources, even
when they are in other rooms .

To reduce this effect a small house should have a shape

for which the mean point-to-point distance is as large as
possible . The mean point-to-point distance is low in compact
shapes like circles and squares , and high in distended shapes
like long thin rectangles .

-115-

11
I ,
P E RIMETER WALL

IN THE PROYECTO EXPERIMENTAL , every lot is surrounded by a

20 cm masonry bearing wall , two storeys high , e x cept in the

back patio where it is one storey high .

THE GENERAL PATTERN IS :

Context:

A house for a low income family , in urban Peru .

Solution:

The house lot is surrounded by a thick masonry wall at

l e ast one storey high .

-116-
Problem:

Peruvians who live in cities are often mistrustful of

their neighbors , and intensely attached to their private prop-
erty - especially to their homes . They feel a tremendous need
for security against the outside world - both against thieves,
and against curious strangers . This feeling is so strong that
many low income Peruvians spend their money on a fine peri-
meter wall of brick or concrete , even when they know this means
they won ' t be able to have a roof , or rooms , inside . It is not
u ncommon to find a family living in a woven bamboo (esteras)
s h ack, inside a masonry perimeter wall . This wall is always
at least one storey high ; in many cases it is also far stronger
than it need be (with extra deep concrete foundation , for ex-
ample} - apparently for psychological reasons , not structural
o n es . (See John Turner , "Lima Barriadas Today ", Architectural
Design , Vol . 33, No . 8 , August 1963 . )

_/.

-117-
CROSS VENTILATED HOUSE

IN THE PROYECTO EXPERIMENTAL , where the pre-

vailing breeze comes from the south , all

houses are oriented north-south . Since each

house has two patios , all living rooms have

openings both south (upwind ) and north (down-

wind) . The southern (upwind ) patio has an

angled scoop to direct breeze into the house .

THE GENERAL PATTERN IS :

Context :

Any building in a hot humid climate, like Lima .

Solution:

Each part of the building is one room thick in the direc-

tion of the breeze , and each room has windows on the upwind

side and on the downwind side .

-118-
Problem:

During the summer months in Lima, the heat is extreme, and

ventilation needed badly. Many people complain about the heat;
few buildings, and especially few small houses, do anything
about it.

The easiest way to make these buildings comfortable, by

non-mechanical means, is to increase the evaporation of mois-
ture by breezes. This means that every room must have two
windows, in line with the prevailing breeze - one upwind, the
other downwind - so that the breeze blows through the room .
(V. Olgyay, Design with Climate, Princeton University Press,
1963, p. 5.)

If these windows open onto patios, the effect of the wind

can be intensified by the use of an air-scoop, in the upwind
patio, to direct the breeze down into the patio. The effect
will also be intensified if the downwind patio is hotter than
the other; when the heat in the hot patio rises, it helps to
pull the air through. This solution has been used for cen-
turies in North Africa and the Middle East .

Since air movement is extremely sensitive to small changes

in profile, the detailed design of the air scoop will depend
on the exact configuration of the patio section, and should be
designed with the help of wind tunnel tests.

I
-119-
11
LIGHT ON T WO SIDES OF EVERY ROOM

IN THE PROYECTO EXPERIMENTAL HOUSE, sala, family room and

kitchen all receive light from both north and south.

THE GENERAL PATTERN IS:

Context:

Any room which is to be used during the daytime .

Solution:

This room is lit, by windows or sky lights, from at least

two directions.

Problem:

A room lit from one side only, is almost always uncomfort-

able. The light gradient on the walls and floors inside the
room is very steep, so that the part furthest from the window
is uncomfortably dark, compared with the part near the window.
Even worse, since there is little reflected light on the room's
inner surfaces, the interior wall immediately next to the win-
dow is usually dark, creating discomfort and glare, against
this light.

Although this glare may be reduced by supplementary arti-

fical lighting, and by well designed window reveals , the most
simple, and most basic way of overcoming gla re, is to give
every room two windows . The light from each wind.ow illuminates
the wall surfaces just inside the other window , thus reducing
the contrast between those walls and the sky outside . For
details, and illustrations, · see R. G. Hopkinson , Architectural
Physics , London HMso·, ·1 963 , pp . 29 and 103 .

-120-
 PATIOS WHICH LIVE

IN THE PROYECTO EXPERIMENTAL HOUSE, both the kitchen patio

and the main living patio are surrounded on three sides by

activities, contain a two meter veranda which connects the

patio to the house, and are placed so that natural circula-

tion moves through the patio.

THE GENERAL PATTERN IS:

Context:

Any patio intended for active use .

Solution:

1. It has sources of traffic and activity on at least two

sides (opposite each other), and functions, at least in

part, as a circulation space.

2. It is placed so that you can see out of it, into some

~ other larger space beyond .

3. At least one side of it is roofed, this roofed part

being at least two meters deep, and connected to the

rest of the building.

-121-

I
j
Problem:

Many of the patios built currently in modern houses are

dead . They are intended to b e private ope n spaces - but often
remain unused .

Informal observation suggests t ha t patios are unused for

the following reasons :

1. No one ever goes t o them when the y do n ot have any natural

rel ation t o the act ivities in the h ouse - thi s is e s p e -
ci al ly tr ue for those that a re d ead-ends , off to one side
of rooms . To overcome thi s , the patio should have activ-
ities , opening off at l eas t two opposite sides , so that
it b e c omes the meeting point to the se activitie s , provide s
access to them, p r ovides overflow from them, and provides
the cross-circulation between them .

2. They are s o enclosed that they become claustrophobic .

Patios which are pleasant to be in always seem to have
"lo opholes " whic h allow yo u to see beyond them , into some

-122-
 further space . The patio should never be perfectly en-
closed by the rooms which surround it , but should give
at least a glimpse, of some other space beyond.

3. They are oppressive . No one wants to sit surrounded by

blank walls, disconnected from the house , with a little
square of sky overhead . To solve this problem, the patio
needs to be partly roofed . This provides a sitting space
that is less nakedly exposed to the sky , and, if the roofed ,I
part is continuous with some interior part of the house, ~·
makes the patio seem more like a part of the house, and
makes it more likely that people will drift naturally into
the patio .

The veranda formed by this overhanging roof will not work

unless it has room for a small table and a couple of chairs ,
so that people can sit there and talk and drink. This re-
quires at least two meters. See "Two Meter Balcony", page 189 .

-123-
TAPESTRY OF LIGHT AND DARK

IN THE PROYECTO EXPERIMENTAL, each house has a sequence of

alternating patios and rooms along its length .

THE GENERAL PATTERN IS:

Context:

Any building where people live during the daytime.

Solution:

Openings and covered areas alternate in such a way that

the interior of the building is a tapestry of alternating

light and dark spaces, with special emphasis on the boundary

areas where dark changes to light.

-124-
Problem:

In a building with uniform light level, there are few

"places" which function as effective settings for human events.
This happens because, to a large extent, the "places" which
make effective settings are defined by light. People are by
nature phototropic - they move towards light, and, when station-
ary, they orient themselves towards the light. As a result the
much loved and much used places in buildings, where the most
things happen, are places like window seats, verandas, fire-
side corners, trellised arbors; all of them defined by non-
uniformities in light, and all of them allowing the people who
are in them to orient themselves towards the light.

There is good reason to believe that people need a rich

variety of settings in their lives (see Roger Barker, The
Structure of Behavior: Explorations of its Structure and I
Content, Appleton-Century-Croft, New York, 1963). Since set-
tings are defined by "places", which in turn seem often to
be defined by light, and since light places can only be de-
fined by contrast with darker ones, this suggests that the
II
interior parts of buildings where people spend much time
should contain a great deal of alternating light and dark.

-125-
PUBLIC PART OF HOUSE

INTIMACY GRADIENT

BATHROOM POSITION

PUERTA FALSA

FIESTA

STAIRCASE STAGE

THICK WALLS
INTIMACY GRADIENT

IN THE PROYECTO EXPERIMENTAL, there is a

strict gradient from formal to informal,

front to back. Each house contains entry-

sala-family room-kitchen in that order.

Those houses too small to have a proper

sala, have a small receiving alcove, just

inside the front door, which functions as

a sala.
..
··---_...<~-- ·-----~:------·~
< ..
( '
kitchen family room sala entry

THE GENERAL PATTERN IS:

Context:

A house in Peru, or any other Latin country.

Solution:

There is a gradient from front to back, from the most

formal at the front, to most intimate and private at the back.

This gradient requires the following strict sequence: Entry-

sala-family room-kitchen-bedrooms .

-128-
 The most important element in this sequence is the sala

(parlor). It is essential that the house contain a sala. If

the house is so small that cost rules this out, the house

should at least contain a tiny receiving alcove immediately

inside the front door.

Problem:

In Latin American countries, such as Peru, friendship is

taken very seriously and exists at a number of levels. Casual
neighborhood friends may never enter one's house. Formal
friends, such as the priest, the daughter's boyfriend and
friends from work may be invited in but tend to be limited to
a well furnished and maintained part of the house, the sala.
This room is sheltered from the clutter and more obvious
poverty of the family which are visible in the rest of the
house. Relatives and intimate friends, such as compadres, may
be made to feel at home in the comedor-estar (family room)
where the family is likely to spend much of its time. A few
relatives and friends, particularly women, will be allowed in-
to the kitchen, other workspaces, and, perhaps, bedrooms of
the house . In this way the family maintains both privacy and
pride.
This is particularly evident at the time of a fiesta. Even
though the house is full of people, some people never get be-
yond the sala ; some don't even get beyond the threshold of the
front door. Others go all the way into the kitchen , where the
cooking is going on, and stay there
throughout the evening . Each per-
son has a very accurate sense of
his degree of intimacy with the
family , and knows exactly how far
into the house he may penetrate ,
according to this established
level of intimacy .

Even e x tremely poor people try

to have a sala if they can . The
photograph shows a sala which a
family has made in a barriada
shack. Yet many modern houses and
apartments in Peru combine sala
and family room in order to save
space. Almost everyone we talk ed
to complained about this situa-
tion . As far as we can tell, a
house must not, under any circum-
stances, violate the principle of
the intimacy gradient.

-129-
BATHROOM POSITION

IN THE PROYECTO EXPERIMENTAL HOUSE, the bathroom is at the

top of the stair in the first patio, directly accessible from

sala, and from bedrooms.

0
n
OL

THE GENERAL PATTERN IS:

Context:

A Peruvian house which contains only one bathroom.

Solution:

The bathroom is placed "between" the public and private

parts of the house, in such a way that the path from sala to

bathroom does not pass through kitchen, family room or bed-

rooms, and the path from bedrooms to bathroom does not pass

through any area visible from the sala or family room.

\
\

Sala Beds

Bathroom

-130-
Problem:

The Peruvian household maintains a strict separation

between the public part of the house (where visitors are re-
ceived) and the family part of the house (which is private).
This is described fully in the "Intimacy Gradient" Pattern
(page 128). If the house has one bathroom, it is very hard
to place this bathroom, so that it does not violate this
separation. I
~'
If the bathroom is in the family part of the house, then ....
a visitor who has to use the bathroom, will be able to see
parts of the house which the family would prefer to keep pri-
vate. If the bathroom is in the public part of the house,
then members of the family will have to go to the public part
of the house, in various states of undress .

. To solve this problem, the bathroom location must satisfy

two requirements:

1. A visitor must be able to get from the sala to the bathroom,

without passing through the family room, kitchen or bed-
rooms, or any other private area.

2. All members of the household must be able to pass between

their bedrooms and the bathroom, without being seen by
anyone in the public part of the house.

These two requirements can be met only if the bathroom is

"between" the public and private parts of the house.

-131-
PUERTA FALSA

IN THE PROYECTO EXPERIMENTAL, 25% of the houses have back

doors. In all houses the path into the house by-passes the

sala.

THE GENERAL PATTERN IS:

Context:

Any Peruvian house.

Solution:

Either the house has a back door, or the main path into

the house by-passes the sala.

Problem:

The puerta f alsa (second entrance) is a crucial problem in

Peru. The nature of Peruvian hospitality, and the formal treat-
ment of strangers, means that while someone is being entertained
in the sala, people must be able to come and go without being
seen.

The women of the house may send the servant, or one of the
children, out for beer, or fruit. In such a case, she does not
want the visitor to realise that she is making a special effort
on his behalf.

Even more simply - the members of the household want to be

able to come and go, without having to involve themselves with
the visitors in the sala, since politeness would then require
them to stay with the visitors and help to entertain them.

-132-
FIESTA

IN THE PROYECTO EXPERIMENTAL HOUSE, the sala, family room,

front patio, and veranda, are distinct, but form a continuous

space where parties can take place.

THE GENERAL PATTERN IS:

Context:

Any Peruvian house.

Solution:

The sala, the family room, and the spaces which connect

them, are distinct, but also form a continuous single space

for parties.

Problem:

In Peru, fiestas (parties) are extremely common. They

happen many times a year, in every household - every time
there is a birthday, name-day, or any kind of special occasion.
Our observations suggest that typical households may have as
many as one a month, and that there may be as many as forty or
fifty people at these fiestas, even in a small house.

This many people can't drink and dance together, unless

the various available living areas (sala, patios, family room,

- '133-
verandas) form a continuous and sizeable space. It is also
necessary that this space be reasonably continuous; the flow
of persons from group to group strongly affects the life of
a party; distinct rooms will hinder this fluidity. On the
other hand, for reasons presented in the "Intimacy Gradient"
pattern the sala and family room must be strictly separated
during everyday use of the house . These rooms must there-
fore be distinct, yet capable of forming a continuous whole
on special occasions.

·~ • I
-134-
 STAIRCASE IS A STAGE

I ' ·\
I I I' '

IN THE PROYECTO EXPERIMENTAL

~ HOU?E, the staircase is in v

-
the frorit patio. It has a

landing on on~' side of the

patio, and the bottom looks t>i 0

0 !> ~ ' -!;:> r

towards the veranda which ~-a, - t:' a-~ !<> "-

connects sala and family " }';!:'?>- .c ~ e ~o '!;;).o a

···_. '.1~~ .\'.S- c 0 0 ~ Cl
room. 1 6 o- . '°' <> Q <Jr /0
). '° 0 ' {> .0 ¢. ! '~ 0
~ ~ b .

THE GENERAL PATTERN IS:

Context:

Any staircase in the

public part of a building.

Solution:

The staircase is connec-

. -~ -.

ted to those places where

social gatherings happen.

The stair is curved if possible, has a wall behind it, is lit

from above and front, and is slightly flared out at the bottom.

The bottom of the stair, and its landings (if any) look out

into the gathering place.

-135-
Problem:

There is little empirical evidence for this pattern. It

is based on the following conjecture. Changes of level play a
crucial role at many moments during social gatherings; they
provide special places to sit, a place where a woman can make
a graceful or dramatic entrance, a place from which to speak,
a place from which to look at other people while also being
seen, a place which increases face to face contact when many
people are together.

If this is so, then the stair is one of the few items in a

building which is capable of providing for this requirement,
since it is one of the few places in a building where a transi-
tion between levels occurs naturally. Stairs should therefore
always be designed to take full advantage of this fact. The
features named in the solution are all intended to increase
the effect of the staircase as a stage.

-136-
 TH IC K WAL LS

IN THE PROYECTO EXPERIMENTAL

HOUSE, the wall connecting

the sala, patio, veranda

and family room, has a ser-

ies of small niches in it,

formed by 40 cm stub walls

that stick out at right

angles to the main wall.

Each niche contains a seat,

I shelves, cupboard or display.

THE GENERAL PATTERN IS:

Context:

Interior wall, in any

part of a building which is

'
intended to be personal.

Solution:

The wall has "depth", at least 40 cm, which is created by

a hand-carvable rigid space frame, in which a continuous variety

of niches, shelves, seats, cupboards, leaning posts, and window

seats occur at frequent intervals. This hand-carvable space

frame is made of materials which are readily available on the

retail market, and easily cut, modified, painted, nailed,

glued, replaced by hand, using only tools available at any

-137-
hardware store. Possible examples are wood, plywood, fiber-

glass styrofoam, polystyrene. The space frame is highly re-

dundant structurally: Large sections of it may be removed,

without weakening it, and pieces or sections may be added in

such a way that these sections become continuous with, and

indistinguishable from, the original surface.

Problem:

Rooms with large, flat, unbroken wall surfaces almost never

have any personal character, and it is very hard for people
who live in such rooms to make them personal. A room becomes
personal, only when the imprint of its inhabitants is clearly
visible, the walls crowded with treasures and belongings
(presents, pictures of sweethearts and grandparents, flowers,
vases, knick-knacks, books, collections), these teasures built
integrally into the fabric of the room, and the surface of the
room moulded to the character of its inhabitants. If a room
has large unbroken wall surfaces, made of unmouldable materials,
none of this is possible. It is hard to store things in the
open, without cluttering up the room, and it is not possible
to build these things in a personal way into the fabric of the
room.

In order to make a room personal, then, its wall surface

must be deep enough to contain a variety of niches and recesses,
where special things can be placed, without being in the way;
and the wall must be made of materials which allow these niches
and recesses to be adapted to the idiosyncracies of the things
which are to be placed there, and to the habits which go with
them. This argument is presented in full, with empirical
evidence, in Christopher Alexander,"Thick Walls", Architectural
Design, February 1968.

-138-
FAMILY PART OF H 0 USE

FAMILY ROOM CIRCULATION

FAMILY ROOM ALCOVES

KITCHEN FAMILY ROOM RELATIONSHIP

HOME WORKSHOP

TWO SERVICE PATIOS

ELBOW ROOM KITCHEN

FAMILY RO .OM CIRCULATION

IN THE PROYECTO EXPERIMENTAL HOUSES , the main part of the

family room {not including a lcoves} , i s 3 - 3.80 meters wide,

running across the lot . There is r oom for a large dining table

in the middle , close to the kitchen .

Circulation from the front of the house

to the kitchen, goes past one end of the

table, perpendicular to the room's main

axis, and there are seats and leaning

niches at this end of the room. The far

end of the room is windowless, and con-

tains an electrical outlet for TV.

THE GENERAL PATTERN IS:

Context:

The family room {comedor-estar} of any low income Peruvian

house.

Solution:

The room is re ;Latively long and narrow. The dining table

is in the middle; traffic in and out of the house goes through

one end, and there are seats or leaning spaces at this end; the

TV set is at the other end, in a darkened corner.

Problem:

For a low income Peruvian family, the family room (comedor-

estar} is the heart of family life . The family eat here, they

-140-
gossip here, they watch TV here, and everyone who comes into
the house comes into this room to say hello to the others, kiss
them, shake hands with them, exchange news, gossip . The same
happens when people leave the house.

The family room cannot function as the heart of the family

life, unless it helps to support these processes. The room
must be so placed in the house, that people naturally pass
through it on their way in and out of the house . The end where
they pass through it, must allow them to linger for a few moments,
without having to pull out a chair to sit down; this requires
"leaning space". The TV set should be at the opposite end of
the room from this throughway; since a glance at the screen, is
often the excuse for a moment's further lingering . If possible
the part of the room for the TV set should be darkened; the
family room, and the TV, function just as much during midday,
as they do at night .

-141-
FAMILY ROOM ALCOVES

IN THE PROYECTO EXPERIMENTAL HOUSES , small family rooms have

one alcove opening off them , and the large family rooms have

two . These alcoves are 250 cm wide and between 120 and 160

cm deep. Ceiling height in the alcoves is 2 . 20 meters (com-

pared with 2.70 in the main part of the family room) .

.~ · ·

-142-
THE GENERAL PATTERN IS:

Context:

The family room of any house.

Solution:

There are a number of al-

coves off the family room (pref-

(D__ J
c
n

)
_)
'---'
erably at least two). Each al-

cove is between one and two

meters deep; the alcoves are

all narrower than the walls they open off; and their ceilings

are lower than the ceiling of the main room. Each alcove

looks at the other alcoves.

Problem:

In modern life, the main function of the family is emo-

tional; it is a source of security and love. But these
qualities will only come into existence, if the members of
the house are physically able to be together, as a family.

In modern life, this is often difficult. The various

members of the family come and go at different times of day;
even when they are in the house, each has his own private
interests: sewing, reading, homework, carpentry, model-
building, games. In many houses, these interests force people
to go off to their own rooms, away from the family. This
happens for two reasons. First, in a normal family room, one
person can easily be disturbed by what the others are doing:
the person who wants to read, is disturbed by the fact that
the others are watching TV. Second, the family room doesn't
usually have any space where people can leave things, and not
have them disturbed. Books left on the dining table, get
cleared away at meal times; a half finished game can't be left
standing; naturally people get into the habit of doing these
things somewhere else - away from the family.

To solve the problem, there must be some way in which the

members of the family can be together, even when they are do-
ing different things. This means that the family room, needs

-143-
a number of small spaces where people can do different things .
The spaces need to be far enough away from the main room , so
that any clutter that develops in them does not encroach on
the communal uses of the main room . The spaces need to be
connected, so that people are still " together" , when they are
in them : this means they need to be open to each other . At
the same time they need to be secluded , so a person in one of
them, is not disturbed by the others . In short , the family
room must be surrounded by small alcoves . The alcoves should
be large enough for one or two people at a time : about two
meters wide , and between one and two meters deep . To make it
clear that they are separate from the main room , so they do
not clutter it up, and so that people in them are secluded ,
they should be narrower than the family room walls , and have
lower ceilings than the main room .

-144-
KITCHEN FAMILY ROOM RELATIONSHIP

IN THE PROYECTO EXPERIMENTAL

HOUSE, the kitchen and family

room open towards one another

across the kitchen patio; kitch-

en counters are hidden behind a

1.30 meter wall; a person work-

ing in the kitchen can see the

family room table and TV.

THE GENERAL PATTERN IS:

Context:

Any house with a family room.

Solution:

The relationship between kitchen and family room follows

two rules:

1. A standing person in the kitchen can see into the family

room, but the kitchen counters cannot be seen from the

family room.

2. Work areas in the kitchen are between 2.5 and 5 meters from

the middle of the family room.

Problem:

The kitchen and family room are always near each other.
Yet the exact relationship between the two poses a subtle con-
flict. On the one hand, people in the family room do not want

-145-
to see too much of the kitchen. This is especially true when
they are eating: they do not want to see food preparation
areas, nor dirty dishes; and if there is a servant in the
house (as there often is in Peru, even in low income house-
holds), they do not want to see the servant in the kitchen.

On the other hand, if the kitchen is too cut off, then the
women of the family will feel isolated, when they are working
there. While they are working, they often want to be in touch
with the rest of the family in the family room, and during the
day they may want to watch the TV, which is probably in the
family room too.

Two features are needed to resolve this conflict. First,

it must be possible for a person standing in the kitchen to
see into the family room, but not possible for a person in the
family room to see the kitchen counters. This can usually be
done by half height walls hiding the counters. Second, the
distance between kitchen work areas and the middle of the fam-
ily room should be between 2.5 and 5 meters. Informal experi-
ments suggest that it is only within this range of distance
that people can be involved, or withdraw from involvement, as
they wish, which is exactly what is needed if the people in
the family room are at some times to feel connected to the
people in the kitchen, and at other times not. These figures
are supported by Edward Hall's figures for "across the room
distance" . According to Hall, at distances between 8 and 20
feet, people can talk, but only by raising their voices.
(Edward Hall., The Silent Language, New York, 1961, p. 164.)

-146-
HOME WORKSHOP

IN THE PROYECTO EXPERIMENTAL , every house contains a family

room alcove that has a view to the front door , and can be used

as a home workshop .

THE GENERAL PATTERN IS :

Contex t :

Any low income house in Peru .

-147-
Solution:

The family part of the house contains space which may be

used as a workshop, with a view to the front door and street.

Problem:

In Peru, many women from low income families take in work

to supplement their incomes: sewing, mending, ironing, small-
time manufacturing. This work needs space.

Wo~e n usually do this kind of work, while watching some-

thing else: something cooking in the kitchen, the TV set, the
children playing. The workspace should therefore be in the
family part of the house. Many women also like to watch the
street while they are working - partly because the children are
likely to be out there, mainly just to be in touch with the
social life. They can only do so, if the workspace commands a
view of the front door and the street.

-148-
TWO SERVICE PATIOS

IN THE PROYECTO EXPERIMENTAL,

every ,house contains two service

patios, both connected to the

kitchen and laundry. One is the

working patio, surrounded by

kitchen and family room. This

patio contains a veranda, and

is covered by sailcloth in win-

ter. The other is the storage patio, for hanging laundry and

storing building materials. I .f extra storage space is needed

for building materials, chickens, etc., a back stair can be

built from the storage patio to the roof.

THE GENERAL PATTERN IS:

Context:

Any low income house in Peru.

Solution:

The house contains two service patios. One is a working

patio: it contains space for table and chairs, and play space

for children, and is connected to the kitchen, laundry and

family room. The other is a storage patio, also connected to

the kitchen and laundry. It cannot be seen from the rest of

the house.
Problem :

Most small Peruvian houses have one service patio only . It

is almost inevitable that this service patio will be used for
hanging laundry , storing building materials and old goods , and
keeping animals . Our observations show few e x ceptions to t h is
rule . Yet these things make the service patio rather unpleasant
to use . The women of the house have to spend a good deal of time
in the service patio , washing clothes and doing housework , and
the smaller children often play near them .

The house needs both : a place for storing things , and a

place for working - but these two places must be separate .
When both are provided, then the storage does not contaminate
the working area; and the working patio becomes a pleasant
place for working and for children's play .

-150-
ELBOW ROOM KITCHEN

IN THE PROYECTO EXPERIMENTAL, families

may choose between two kitchens: a small

one and a large one. The large one has

6 meters of counter and room for a kitch-

en table and corner seat.

THE GENERAL PATTERN IS:

Context:

Any Peruvian house for large low income families.

Solution:

The kitchen is large enough to contain a kitchen table,

and at least 3.60 meters of counter.

Problem:

In a Peruvian household, the kitchen is often used by

several people at once. This is especially important during
a fiesta, when all the women of the family will crowd into the
kitchen to help prepare food and serve guests.

At such times, or if anyone is trying to eat in the

kitchen, there must be plenty of room in the kitchen - at
least room for a table, and room for three people working
(3.60 meters, at 1.20 meters each).

-151-
SLEEPING AREAS

INDIVIDUAL BED ALCOVES

BED CLUSTERS

MASTER BEDROOM LOCATION

MASTER BEDROOM DRESSING SPACES

OLD PEOPLE DOWNSTAIRS

SERVANT SLEEPING SPACE

TWO COMPARTMENT BATHROOM

CLOTHES DRYING CLOSET

INDIVIDUAL BED ALCOVES

THE PROYECTO EXPERIMENTAL

HOUSE, contains a number of

curtained alcoves , each large

enough for a bed and storage .

There is one alcove for each

child - or small children

may share an alcove with

bunk beds . Each alcove opens

o ff a common space for play

a nd circulation, which it-

self opens off a patio.

THE GENERAL PATTERN IS:

Context:

A small house with many

children.

Solution:

Each child in the house gets a private, curtained alcove,

at least 2.00 meters long, and at least 1.30 meters deep.

Each alcove contains a bed, and storage space; it is enclosed

on three sides, and open to a larger space on the long side

opposite the bed, with a curtain to close this long side.

-154-
Problem:

In any small house with many people in it, individuals get

very little privacy. In Peru, for instance, where families
are particularly overcrowded, people complain constantly about
the lack of privacy. This lack of privacy centers especi~lly
on two issues:

1. Every person in a household needs at least one tiny place

which he (or she) can call his own, where he and his things
will be undisturbed.

2. People need a place where they can dress and undress in

private. Many Peruvians, especially girls, are extremely
modest. In a small house, where many people share bedrooms,
they often have to dress and undress in the bathroom.

With a large family in a small house, it is obviously not

possible to give each child a room of his own. (In fact, it
is not even desirable to do so - especially in a country like
Peru~ where people are uncomfortable when isolated. This is
discussed fully in "Bed Clusters", page 156.) The use of small
curtained bed alcoves, opening off larger common spaces, is a
way of giving each person a reasonable amount of privacy, with-
out using an unreasonable amount of space to do it. In 2.00 x
1.30 meters, there is room for a bed, for clothes, for dress-
ing and undressing, and for personal belongings and decorations.

-155-
BED CLUSTERS

IN THE PROYECTO EXPERIMENTAL HOUSE, there are two clusters of

bed alcoves - one around the front patio, the other around the

second patio. Each may have up to five beds in it.

0
0
0
OD

THE GENERAL PATTERN IS:

Context:

The sleeping areas of a Peruvian house.

Solution:

The children's beds are arranged around common areas, to

form strongly inward looking clusters. There are at least two

distinct clusters, one for boys and one for girls.

Problem:

Whether the house contains beds, or bed alcoves (as dis-

cussed under Individual Bed Alcoves), it is very important
that the beds be grouped in clusters.

Peruvians do not like to feel isolated, and draw a great

deal of comfort and security from the fact that they are con-
stantly surrounded by people. It is considered impolite to
leave someone alone. They are also unusually afraid of bur-
glars and robbers, and need to feel very secure where they
sleep. It is therefore necessary that people sleep "in the
presence of others". If the beds are in bed alcoves, then
these alcoves must be arranged in a strongly inward looking
manner.

The house must contain at least two clusters of beds. In

Peru, parents insist on strong segregation of the sexes in the
sleeping areas of the house - except for the tiniest children.
There should be at least one cluster for boys, and one for
girls.

-156-
MASTER BEDROOM LOCATION

IN THE PROYECTO EXPERIMENTAL HOUSE, the master bedroom is at

the top of the stair. It has a window overlooking the inside

of the front door, has direct access to the bathroom; and has

no common walls with any other rooms in the house.

THE GENERAL PATTERN IS:

Context:

Any low income Peruvian family house.

Solution:

The master bedroom has a view of the front door, is at the

head of the stair (in a two-storey house), is next to a bath-

room, and shares no walls with other rooms.

Problem:

The master bedroom's locational requirements are:

1. Control over children's coming and going - this means the

bedroom must be at the head of the stair (in a two-storey
house), with a view of the front door.

2. Fear of burglars and robbers - again, control of the front

door, so that it is possible to check anyone who comes into
the house.

-157-
3. Direct access to the bathroom.

4. Maximwn acoustic privacy - in a cheap house, where par-

titions are cheap and insubstantial, this means the
master bedroom should not share walls with any other
inhabited rooms.

-158-
MASTER BEDROOM DRESSING SPACES

IN THE PROYECTO EXPERIMENTAL

HOUSE , t he master b edr o om has ,

at one e nd , a sma ll curtained

. I
d ress i ng a r ea o pening off it , . .'
....
.. ,
·:·,.
. . !.:
~ . ~ '.;
/
large enou gh t o c ontain two .
~~··
·
··~===.~ ..
' .' . · ' '
separate close t s , a nd larg e

eno u g h to hold a baby ' s crib .

We h a v e no t b een a ble to in-

elude t h e p atter n in its proper

f o rm .

THE GENERAL PATTERN IS : ==- - .

o__ ~

Context :
-
A master bedroom in any house.

Solution:

The room contains two dressing spaces, one for the man ,

one for the woman, each with its own closets. The woman's is

slightly larger than the man ' s, and large enough to hold a

baby's crib. Under ideal circumstances, these dressing spaces

should be tiny rooms, opening off the bedroom.

Problem:

A master bedroom needs two separate dressing spaces in it.

Man and wife of ten disagree in small ways about the use of the
master bedroom , whether clothes should be hung up in a closet

-159-

l
or over a chair, and other basically trivial, but potentially
annoying issues. These issues reflect the more important
fact that each person needs at least a small part of the room
set aside for his own things, and his own clothes. In a small
house these areas cannot be more than special dressing alcoves.
In a larger house, the ideal bedroom has two dressing rooms
attached to it, each large enough to take on the full identity
of the person who uses it.

Since women spend far more time in the master bedroom than
men do, this is particularly critical for the woman. Her
dressing place should be large enough for ornaments and personal
belongings, and large enough to contain a crib for a new born
baby, or a chair where she can sew.

-160-
OLD PEOPLE DOWNSTAIRS

IN THE PROYECTO EXPERIMENTAL HOUSE, a bed alcove for a grand-

mother or grandfather may be built immediately next to the

family room, in the kitchen patio. If a toilet is to be built

with it, this may be placed in the storage patio, where ~t can

be connected to the plumbing wall.

THE GENERAL PATTERN IS :

Context:

Any low income house in Peru.

Solution:

The ground floor of the house contains a place where a bed-

room or bed alcove might be built . It should be close t o the

family room, c l ose to kitchen and laundry , and reasonably close

to a ground floor bathroom (o r place where a ground floor bath-

room might be built ) •

Problem :

Many low income households in Peru contain an elderly re l a-

tive. For example, in 195 6 , 26 % of barriada households had a
relative who was not a member of the nuclear family , living in .
(Jose Matos Mar, "Migration and Urbanisation", in Urbanisat i o n

-161-
in Latin America, International Documents Service, Columbia
University Press, New York, 1961, p. 181.)

The space provided for an old person needs special char-

acteristics. First, many old people have trouble with stairs.
In a two storey house, their room should be on the ground
floor, with the possiblity of a bathroom on the ground floor
too. Second, many old people are bedridden. They need con-
stant care and entertainment; but if they are stuck in a bed-
room far away from living areas, they often get left alone.
The old person's bedroom should be near the kitchen and
laundry, where much of the daily work goes on, and also near
the family room, so that family social activity, and TV watch-
ing, are close by.

-162-
SERVANT SLEEPING SPACE

IN THE PROYECTO EXPERIMENTAL

HOUSE, there is room for a ser-

r
vant to sleep far from the rest

of the house, either in the back

storage patio or on the roof

with access from the patio by

ladder-stair.

THE GENERAL PATTERN IS:

Context:

Low income Peruvian house.

Solution:

There is room for easy construction of servant sleeping

space, as far removed as possible from the rest of the house.

Problem:

In Peru even relatively poor households will often have

servants. Many families will not be able to afford special
accommodations for the servant,
but will allow the servant to
sleep on the roof or in a back
patio if the servant has no
family near by.

Since the servant will not

be accepted into the family cir-
cle, he (or she) needs a private
place of his own. This is es-
pecially important for those
servants of school age (very
common in Lima) , who need a
place to study; often the
chance to study is the only
chance these young people have
of improving their lives.

-163-
TWO COMPARTMENT BATHROOM

IN THE PROYECTO EXPERIMENTAL, even the cheapest house contains

two bathroom compartments. One compartment contains the toilet,

the other the wash basin and shower.

THE GENERAL PATTERN IS:

Context:

Any house which is to contain many people, and has only

one bathroom.

Solution:

The bathroom has at least two compartments: toilet in

one, wash basin and shower in the other.

Problem:

In large families with one bathroom, there is a major queu-

ing problem in the morning. Everyone wants to use the toilet,
wash basin, and shower at the same time. Separate, their use
is doubled.

-164-
CLOTHES DRYING CLOSET

IN THE PROYECTO EXPERIMENTAL HOUSE, a two square meter closet

with drying racks and an electrical outlet for a space heater

opens into the shower room and is ventilated by a skylight.

THE GENERAL PATTERN IS:

Context:

Any house in Lima without a heating system.

Solution:

The house contains a larg~ well ventilated closet with

open hanging racks, and a heater.

Problem:

In any cool humid climate, like the Lima winter, laundry

does not dry fast enough in the open, and it is very hard to
keep the family supplied with warm dry clothes. The house
should contain a clothes drying closet, with open hanging
racks, near to some source of heat. In Lima, many families
in small unheated houses buy space heaters and put them in the
bathroom: the drying closet should open directly into the
bathroom.

-165-
ENTRANCE AND FACADE

ENTRANCE TRANSITION

FRONT DOOR RECESSES

MIRADOR

FRONT DOOR BENCH

GALLERY SURROUND

NO GROUND FLOOR WINDOWS ON THE STREET

ENTRANCE TRANSITION

IN THE PROYECTO EXPERIMENTAL HOUSE, there is a dark, covered

area, immediately· inside the front door, and a well lit patio

further in. A person entering thus passes through a dark

zone, towards the light of the patio beyond, and then enters

the house through the main veranda, between sala and family

room.

THE GENERAL PATTERN IS :

Context :

Any house entrance .

Solution :

The path from the street

into the house passes through a

zone where levels , materials ,

view , light and other qualities

change .

-168-
Problem :

There is no doubt that houses which provide a graceful

transition space between street and house , are nicer than those
which open abruptly off the street . If the transition is too
abrupt there is no feeling of arrival , and the house .fails to
be an inner sanctum.

The following argument may help to explain it . While

people are on the street , they adopt a mask of "street behavior" .
When they come into a house they naturally want to get rid of
this street behavior and settle down completely , into the more
intimate spirit appropriate to a house . But . it seems likely
that they cannot do this , unless there i~ a tran~ition from one
to the other , which helps them to lose the street behavior .
The transition must , in effect , destroy the momentum of the
closedness , tension and "distance" which are ap~ropriate to
street behavior , before they can relax completely. ·.
' '

Evidence comes from the report by S~rge Bout:eriine and

Robert Weiss , The Seattle World ' s Fair , Cambridge , Mass. , 1963 .
The authors noticed that many exhibits failed ..to· "hold" people;
people drifted in , and then drifted orit agairi ~i£hin a very
short time . However , in one exhibit people h~d to cross a
huge , deep-pile , bright orange carpet on the .way in : in this
case , though the exhibit was no better than other exhibits ,
people stayed . The authors concluded that people were , in
genera l , under the influence of their own "street and crowd
behaviour ", and that while under this influence could not
relax enough to make contact with the exhibits ; but that the
bright carpet presented them with such a strong contrast as
they walked in , that it broke the effect of their outside
behavior , in effect "wiped them clean" , with the result that
they could then get absorbed in the exhibit.

There are many ways of marking the transition from street

to house : change of view, change of light , change of level ,
change of surface, change of sound, change of scale , all break
the continuity of passage from street to house, and can all be
helpful .

-169-
FRONT DOOR RECESSES

IN THE PROYECTO EXPERIMENTAL

HOUSE, each front door is j

~I
surrounded on the outside by 'I

one or more deep recesses,

according to the exact posi-

tion of the entrance with

respect to other houses.

The front doors are of the

0
dutch door type. 7 -----

~-- -· -

THE GENERAL PATTERN IS:

Context:

Any Peruvian house which

has a front door opening '

\
directly off a public path.

Solution:

The front door is surrounded, on both sides, by deep

recesses, each at least SO cm deep - if possible by double

recesses. The effect of the recesses is helped by an open-

ing in the door, or a dutch door.

-170-
Problem :

"Hanging out " is a standard part of Latin culture . People

like to watch the street . But people do not always want the
same degree of involvement with the street. The process of
hanging out requires a continuum of degrees of involvement
with the street , ranging all the way from the most private
kind, to the most public kind . A young girl watching the

-171-
street may want to be able to withdraw the moment anyone looks
at here too intently. At other times, girls, young men, and
the women of the house, may want to be watching the street,
near enough to it to talk to someone who comes past, yet still
protected enough so that they can withdraw into their own
domain at a moment's notice. At still other times, old men,
less afraid of real involvement on the street, will actually
sit out, in front of their doorways, and feel secure provided
that the seat is still clearly identified with their house.

In the most common kind of hanging out, people lean in door-

ways, half in, half out. They can see what is happening out-
side, they can talk to anyone they want to - yet they can
withdraw in a moment. To invite thjs activity, front doors
need deep recesses, large enough to hold a person (thus at
least 50 cm deep), and, if possible, a way of hanging over the
door, like that which a dutch door provides. The other two
kinds of hanging out are discussed under "Mirador" (page 173) ,
and "Front Door Bench" (page 175).

-172-
MIRADOR

IN THE PROYECTO EXPERIMENTAL , a second floor window opens from

the front bed cluster , onto the pedestrian street below .

Corner houses also have a mirador on the side wall, at the top

of the stair .

I
I
~1

l -!
1
i
l1
j
THE GENERAL PATTERN IS:

Context:

Any house in Peru.

Solution:

A second floor window , with a seat by

it , looks up and down the street outside .

This window is either in the girl ' s bed -

room, or on the passage between this

bedroom and the stair .

Problem :

The most private kind of involvement

with the street is watching from an upper
storey window . For young girls , espec-
ially , this is a favorite activity - they
can watch the street , from a window , with-
out any impropriety : something they
cannot do so easily from the front door .
If anyone looks at them too hard , they
can pull b ack into the window .

To be tru ly useful , such a window

must be closely associated with places
where the girls will often be - either
the girls ' bedroom , or the passage be -
tween this bedroom and the stair .

The process of watching the stree t

from upper storey windows is strongly
embedded in traditional Peruvian culture ,
in the form of the "mirador ", the
beautiful ornamented gallery which sticks
out over the street from many colonial
buildings in Lima.
FRONT DOOR BENCH

IN THE PROYECTO EXPERIMENTAL, houses have some kind of bench

outside the front door, placed so as to create a private space

just in front of the door . The benches will vary in position,

size, and coloring, from house to house.

I I
I !

_- -:-:·- - .. _.:..:: - - .- - . - - .

THE GENERAL PATTERN IS:

Context:
(

Any Peruvian house with a front door opening directly off

the public path.

Solution:

There is a bench outside the front door. This bench may

be directly in front of the door, or to one side, but is

placed to create a small private space immediately in front

of the door.

-175-
Problem :

The most public kind of involvement with the street is

sitting out. . (See "Front Door Recesses" for general statement .)
Many people, especially older people, pull chairs out to the
front door , or lean against the front of their houses , either
while they are working at something , or just for the pleasure
of watching street life . Since people in Peru are reluctant
to be too public 7 this activity requir~s a bench or seat whic h
is clearly private , even though i n the public world .

I
•

-176-
GALLERY SURROUND

IN THE PROYECTO EXPERIMENTAL, each house has an open gallery

at the front of the second storey. In the smaller houses,

this gallery is very narrow , and runs back into the front bed

cluster; in the larger houses, it runs all the way across the
house.

r;
I (
I
I

) '

I
j I
I
·1

! ,
l/
/

THE GENERAL PATTERN IS:

Context:

Mass housing.

-177-
Solution:

Each house has an open gallery, at least two meters deep,

on its visible sides: This gallery is intended to be filled

with furniture, extensions, partitions, plants, seats, decor-

ations.

Problem:

One of the worst features of mass housing is the uniform-

ity of the houses. People want to live in houses which are
individual, identifiable, personal. Paint and color and decor-
ation don't do it. A house which has nothing but walls and
windows on the street, will almost certainly remain anonymous,
even many years after it is inhabited. What is needed is some
outside to the house, which is enough "part of the inside" so
that it will quickly be made personal and different from the
neighbors. This requires a facade which is essentially "deep"

and able to grow and change with time - an unfinished gallery

round the house. After a few years, each gallery will contain
a unique collection of extensions, planting, furnishing, and
decoration - and will be quite different from its neighbors.

The argument is essentially the same as that presented under

"Thick Walls", applied to the outside of the house. See page
137. The reasons for the gallery to be two meters deep are
given under "Two Meter Balcony", page 189.

-178-
NO GROUND FLOOR WINDOWS ON
THE STREET

THE PROYECTO EXPERIMENTAL HOUSE, has no

ground floor windows onto the street.

The sala does get light from the street,

but through a horizontal slit at ceiling

level, baffled by the beam.

THE GENERAL PATTERN IS:

Context:

Any Peruvian house.

Solution:

The wall facing the street at ground floor level, is

blank, except for the door. There are no windows in it.

Problem:

When Peruvian houses have windows facing onto the street

at ground level, these windows are always boarded up, painted
with whitewash, heavily curtained and screened. A survey of
60 houses in San Martin de Porras (a part of Lima), showed
only two first floor windows which were not completely ob-
scured by curtains, shutters, blinds, paint, or other opaque
materials. One of the two opened onto a home workshop; the
other into a shop. People are extremely private. They do
not want passers-by to see into their house, and want secure
protection against thieves.

-179-
 It is important to take this into account from the very
outset, in the way that front rooms are lit and ventilated.
If the front room relies on a street window for light and air,
the inevitable boa~ding up will make the room dark and stuffy ,
and less usable.

- 180-
PAT I 0 SECTION

TRANSLUCENT OPENING PATIO ROOF

LIGHT FROM TWO STOREY PATIOS

SUNSHINE IN PATIOS

TWO METER BALCONY

TRANSLUCENT OPENING PAT I 0 ROOF

THE PROYECTO EXPERIMENTAL HOUSE, can be completely sealed

against wind, cold, and dew, by dacron sailcloth covers on

the front patio and kitchen patio. The dacron sailcloth is

translucent, and does not stretch or mildew. It is fastened

to the roof on one side, and has snap fasteners on the oppo-

site side. On the other two sides, it has rings sewn into

it: these rings run on a pair of parallel rods. A person

can open and close the cover, or set it to any desired opening,

by pulling cords just like those that control conventional

draw-curtains. The sailcloth has wires sewn into it at one

-182-
meter intervals, perpendicular to the rods, thus creating a

tension structure which resists wind loads and prevents tear-

ing and sagging.

Since the house can be completely sealed by this cover,

the windows which open into patios have no glass or moving

frames in them. For visual privacy, the windows can be closed

by curtains.

THE GENERAL PATTERN IS:

Context:

Any patio house, in a warm climate without rain.

Solution:

Patios are covered with a translucent lightweight cover,

which can be rolled back in hot weather, and closed in cool

weather . The windows which open into these patios need no

glass, only screens or curtains.

Problem:

In a warm climate, houses need large openings in summer,

and good protection against cold in winter and at night . Yet
windows are very expensive. In many houses, especially in low
cost houses, the cost of the windows makes it impossible to
create the proper relationship between indoors and outdoors .

As a result, the house is unresponsive to the year-round

change in climate. This is particularly serious in low cost
houses. Newly built low cost houses in Lima, for example, are
so tightly budgeted that they often have far too little window
area: they are unbearably hot in summer, and dark and gloomy
all year round.

-183-
 It is possible to solve this problem, in a patio house,
by roofing the patios with movable, translucent covers. With
such covers, the whole house - patios and rooms together -
becomes a sealable envelope. Climate control, instead of in-
cluding rooms only, includes the whole house , inside and out.

In hot weather, the patio covers can be rolled back, and

the breeze blows through the house. In cold weather, the patio
covers are closed, thus sealing the house - and the patios be-
come usable too. In Lima, for example, there is a winter dew
(garua) which normally makes patio floors damp and cold for
eight months in the year. The cover on the patios will keep
them dry and warm, and will triple their useful life.

If patios are small, the patio roofs can be made of low

cost materials like the dacron sailcloth described above . They
eliminate the need for windows almost entirely. The windows
which look into patios give light to rooms and may be curtained,
for visual control - but since the cold is kept out by the patio
covers, there need be no glass in the windows , and no expensive
moving parts.

-184-
LIGHT FROM TWO STOREY PATIOS

IN THE 'PROYECTO EXPERIMENTAL HOUSE, the front patio opening

is 2.90 meters by 4.40 meters or more; the kitchen patio open-

ing is 2.90 meters by 2.40 meters. Both these openings are

slightly more than half the area of their respective patios.

All space inside the house is within 2 meters of a patio.

THE GENERAL PATTERN IS:

Context:

A two storey patio house in Lima.

Solution:

The patios follow three rules:

1. No part of any room is more than 2 meters from an opening

into a patio.

2. In each patio the opening to the sky is at least 2.50 meters

in each direction .

3. The patio opening is substantially smaller than the patio

itself .

-185-
Problem:

In a two storey house, if patios are too small, the rooms

are dark and gloomy; if the patios are too open, the light is
harsh and unpleasant. It is almost impossible to calculate
this effect by means of daylight factors - we therefore made
a number of observations to isolate the shapes and sizes of
patios which create a pleasant light in the rooms which open
off them. These observations were all made in Lima, and depend
on the particular sky brightness and light quality characteristic
of Lima. It is possible that similar figures apply in other
places - but the light in Lima has a fairly unusual haze .

We found that any part of a room that is more than 2 meters

from a two storey patio, is gloomy; that in general, the room
will be gloomy, if the patio opening is less than 2.50 meters
in any direction; and that the light will be too harsh, if the
patio opening is not substantially smaller than the patio itself.

-186-
SUNSHINE IN PATIOS

IN THE PROYECTO EXPERIMENTAL , houses have a 30 cm overhang

protecting all first floor openings. There gre no openings

facing west . The main patio is always at least 4 . 40 meters

long in the north-south direction . Either sala or family room

always gets north (i . e ., winter) sun from this main patio.

THE GENERAL PATTERN IS :

Context:

Any pat io hou se in Lima .

Solution :

Three rules apply :

1. All north and south fa c ing

openings are pro t e c ted by a n

overhang which is 21% of the

height from window si l l t o

overhang .

2. No opening is exposed t o the we st .

3. One patio is long in the north-sou t h direc t ion - its length

at least 7 3% of the hei ght o f t h e north wall - and one l iv-

ing r oom faces no r t h i nto th i s pa tio.

Problem :

During the s ummer in Lima , the ove rhe ad sun is fie rce , and
rooms mu s t b e prote cte d f rom it . During summe r months , t h e s ~n
go e s no rth and s outh , but n e v e r goe s b e low 78 d egrees on the

-187-
 Zenith
Equinox Winter
Summer Solstice
Solstice

SUN ANGLES

north-south axis. At this angle, overhangs must be 21% of

the height of the wall which is to be protected (tan 12° =
.21), to keep the wall in shade.

The late afternoon sun is also extremely hot; it is hard

to shade against it, and best to avoid openings which face
west.

In winter there is very little sun. People enjoy sitting

in what sun there is, and, if possible, like to get it into
living areas. In mid-winter, the sun is low in the north
(54 degrees). At this angle, the north-south length of a
patio must be at least 73% of the height of the north wall
(tan 36° = .73), to get sun on the ground.

-188-
TWO METER BALCONY

IN THE PROYECTO EXPERIMENTAL, each house contains a two meter

deep v~randa, both in the main patio and in the kitchen patio;

the upstairs front gallery, which looks onto the street, is

two meters deep; and the arcades which surround the outdoor

rooms, outside the houses, are always at least two meters deep.

THE GENERAL PATTERN IS:

Context:

Any habitable indoor-outdoor space, like a balcony, veranda,

gallery, porch, deck or arcade.

Solution:

It must be at least two meters deep.

-189-
Problem :

Balconies and porches which are less than two meters deep ,
are almost never used . They are often made small, to save
money ; but when they are small , they might just as well not
be there . They are first used properly, when there is enough
room for two or three people to sit there in a small group ,
with room to stretch their legs , and room for a small table
where they can put glasses , cups , and so on . No balcony works
if it is so narrow, that people have to sit in a row , facing
outwards.

The critical size is hard to determine exactly . It is

about two meters .

-190-
SHOP AND RENTAL

SHOP FRONT POSSIBILITY

RENTAL
SHOP FRONT POSSIBILITY

IN THE PROYECTO EXPERIMENTAL, families who want a shop now or

the possibility of a shop in the future, get a house in which

there is space for a shop on a suitable exterior corner of the

house.
Medium or
Tiny Sala Large Sala

Shop Facing Market

Medium or Medium or
Tiny Sala Tiny Sala Larqe Sala Larqe Sala
West Corner

b.· ~.I~
Ea.at Corner Eaa t Corner West Corner

I t~~;
, . . . - · · .

Shop at Front of House

lk~ ~
&u.ll Kitchen S.all Kitchen Larqe Kitchen Larqe Xi tchen
Eas t corner West Corner East Corner West Corner

Shop at Back of House

THE GENERAL PATTERN IS:

Context:

House for a low income Peruvian family.

Solution:

The house contains space where the owners can start a smal l

shop if they want to. The space need be no larger than 2 x 3

meters and should be in sight of the family room or kitchen.

-192-
Prob lem :

Many families with low incomes , start a small s hop t o

supplement the i r income . Figures f or l ow to mo derate income
sectio ns of Lima , for e x ample , i n di cate t hat 10% o f the house s
contain some type of shop o r workshop . (Jos e Matos Mar,
"Migration and Urbanization" , in Urb anizatio n in Latin Ameri ca ,
e dite d by Phillip Hauser , Int ernatio nal Documents Se rvi ce ,
Columbi a Unive rsity Press and UNESCO , 1961 , page 18 1 .) Our own
count gave the same figur e . We s uspe ct that g ove rnme nt poli cy
may not allow the construction o f sma ll shops in public hous-
ing . However , a sma ll surv e y o f empleados showed that many of
them desire this option , and would in many cases prefer it to
the provisio n of othe r kinds of space . We strong ly urge that
they be allowed to build stor es if they want to .

From a physical point of view , the problem is rather simple .

The house shorild containJ space , at least 2 x 3 meters, on a n
exterior fac~ of the building , preferably on the corner where
the most people are going past . This space must be placed so
that the inside of the house will still be private and secure ,
if there is a shop built there; and if possible , it should be
visible from family room or kitchen , so that members of the
family can keep an eye on it from the family part of the house .

-193-
RENTAL

IN THE PROYECTO EXPERIMENTAL, fam- Open on

Back.
Open on
west
Open on
East

~~
~q ~q •Kfj
ilies who want the opportunity to

~~
rent a room now or later, get a cor-
Small Kitchen

ner lot or a lot which is free in

_. . . ~ ~ g;J
the back, so that they can build

rentable space there .

THE GENERAL PATTERN IS:

Context:

Any low income household in Peru.

Solution:

The house contains potentially rentable space, with the

following characteristics:

1. The space is between 6 and 10 square meters.

2. It has a separate entry.

3. It has no association with the formal part of the house

(sala, front door, etc.).

Problem:

It is common for Peruvian families to rent space - even

when it is not officially allowed - as a source of extra
income.

The usual renters are either single, or couples without

children - they need no more than one small room (6-10 square
meters). They do not need a special bathroom, and will either
share the facilities the servant uses, or will use a chamber
pot, pitcher and enamel wash basin.

Families also naturally want to keep their distance from

the people they are renting to; they will try to avoid having
the renter come and go through the main part of the house, and
much prefer it if they can give him a separate entrance, in
the back of the house.

-194-
CONSTRUCT .ION

CONTINUOUS FLOATING SLAB

MORTARLESS BLOCK WALL

COMPOSITE BAMBOO/FOAM BEAM

COMPOSITE BAMBOO/FOAM PLANK

SULPHUR REINFORCING AND TOPPING

PLUMBING ACCUMULATOR

CONTINUOUS ELECTRIC OUTLET

 The following seven patterns describe building components

or appliances which have a distinctly innovative character, and

the contexts in which these innovations are appropriate. Most

of them are extensions of traditional technology currently in

use in Peru.

Our designs also include several building components that

are standard items now in wide use in Peru: gypsum partitions,

wooden doors, bathroom fixtures, plastic piping, etc. Since

these items are well known, we have not given any general

patterns for them.

In addition, we are using several standard methods of site

development with an emphasis on economy. The internal loop

roads will have the minimum amount of paving consisting of

two narrow tracks of concrete or asphalt paving with unfinished

earth between. Street lights will be strung on cables between

buildings, except where buildings or column location allow

direct mounting of fixtures on them. Outdoor public rooms

(page 102) will be made from concrete block columns supporting

bamboo beams and woven bamboo mats in frames. Utility lines

will run in trenches down pedestrian paths with laterals to

individual buildings - the laterals are as short as possible.

-196-
CONTINUOUS FLOATING SL A B

IN THE PROYECTO EXPERIMENTAL, the

floors of all houses are 10 cm con-

crete slabs laid down in 5 . 20 meter

strips (width of houses ) by a road

building machine known as a slip-form

paver . All houses are parallel . The

2
slab is 175 kg/cm concrete on 30 cm

of compacted sand . All e x posed edges

are turned down , filled block (20 cm

x 20 cm ). Structural pin connections

for ties into walls are cast by hand immediately after the

slab is laid. The slab is dustcoated and colored with a deep

red penetrating oil stain for finish .

Pedestrian walkways and patio floors in the houses are also

cast in the same way, but onto prelaid dividers to form 50 cm x

50 cm paving stones inside and 1 meter square paving stones

outside.

Slab Edge Wall Tie

THE GENERAL PATTERN IS:

Context:

Mass housing with slab on grade, on a flat site and on

soil with normal bearing capacity.

Solution:

The floors are unreinforced concrete slabs with no foot-

ings or foundations, laid down by a road building machine.

Slab thickness varies according to exact site conditions and

building loads.

Slabs should be laid out in continuous strips of constant

width, at least 50 meters long. Joints and breaks may be in-

serted as required. If the slab is laid on a curve, the

radius of the curve is not less than 20 meters.

Problem:

Foundations, footings and floors are a major cost in house

construction. Foundations and footings usually represent
about 10% of the total construction cost, and floor construc-
tion another 10%. To eliminate individually excavated and
poured foundations, it is possible to use a continuous float-
ing slab, produced by a road building machine, at great speed
and low cost. Vertical loads are distributed uniformly over
the slab and make footings unnecessary: horizontally, the
slab is held by a turndown at the exposee edges - made by pre-
trenching. If the slab needs reinforcement, mesh is laid by
hand in front of the machine. However, in warm climates like
Peru, there is little temperature change in the soil, and slabs
need no reinforcement. Pins which tie the building to the slab
are put into the slab by hand, while it is wet. The system
is highly earthquake resistant, since the slab and the ' build-
ing act ,as a unit, floating on the earth, instead of being tied
to the earth by footings and foundations.

Road building machines are available in almost all countries

today - wherever high speed roads are being built. In order to
use these machines efficiently, the site must be reasonably
flat, and a large number of slabs must be laid at the same time.
Furthermore, the slabs must be laid in long strips of constant
width, adjacent strips must be parallel, and the strips may
have a minimum radius no less than 20 meters (the smallest curve
these machines can negotiate).

-199-
MORTAR LESS BLOCK WALL

IN THE PROYECTO EXPERIMENTAL , the bearing walls , shear walls ,

and ground floor partitions are cavity walls made of inter-

locking , self-aligning concrete block s , moulded o n si t e . No

mortar is required . The blocks simp ly i nte rloc k to fo r m the

wa ll . Blo cks a r e made of a dry c o ncrete mi x, a nd may b e 1 0

or 20 cm thick: column and corner blocks are made in specia l

moulds . Walls and columns are reinforced with sulphur (see

page 214). Plumbing lines and electrical conduits run through

the cavity.

-200-
THE GENERAL PATTERN IS:

Context:

Low cost, low rise building, in any place where concrete

is one of the cheapest building materials.

Solution:

Bearing walls, shear walls, columns, partition walls, and

foundations may be made from EDI-Thermomod blocks. The block

-201-
is self supporting up to a height of three stories . Vertical

edges, corners, and horizontal upper edges are either poured

concrete, with steel reinforcing, or are reinforced with

sulphur-fiber. The EDI-Thermomod system is patented by

Educational Design, Inc .

'
'' ''
/ /
/ /
_/

Half Blocks

Problem:

In areas where concrete is the principal building material ,

concrete block is one of the cheapest forms of wall construc-
tion . One of the biggest costs of a concrete block wall is the
labor cost: each block has to be placed and mortared by a
skilled mason. The EDI-Thermomod block system eliminates al-
most all these labor qosts . The blocks are mortarless and
self aligning; they weigh only 5 kg apiece, and are very easy
to handle. Two men can build a wall extremely fast, simply
by stacking the blocks on one another . Masons are not required .

Another major cost in concrete block construction is the

cost of the block itself . Here again the EDI-Thermomod block
saves money. The block can be hand-manufactured on site in a
simple mould, or machine manufactured . One mould produces
about 400 blocks in 8 hours; a battery of five moulds will pro-
duce about 2,000 blocks a day - enough for the walls of an

-202-
average 100 square meter house. Half blocks and blocks of
different thickness can be made from the same mould. The
blocks are cured after 24 hours. On site manufacture elimin-
ates expensive storage and trucking.

The wall has several other advantages: The system is

light, and earthquake resistant. The dry construction allows
the blocks to move during a quake, thus preventing fracture
of the wall. Since there is no mortar, blocks can be removed
at any time to make new openings in the wall. The cavity can
be made to serve as a conduit for plubming and electricity -
because the blocks can be removed, the conduits are easy to
reach. Finally, like any cavity wall, the wall has good ther-
mal and acoustic insulation.

The EDI block has been used to build very low cost build-
ings in Mexico and the south-eastern United States. Many of
these buildings were built entirely by self-help; the build-
ings are performing well in use; those in Mexico have success-
fully withstood major earthquakes.

-203-
COMPOSITE BAMBOO FOAM BEAM

IN THE PROYECTO EXPERIMENTAL HOUSE, all

beams are rectangular section boxed

beams, 20 cm wide, 40 cm deep, and 5

meters long. The beams are made of 6 cm

bamboos, placed over plywood templates,

with a core of two lb. density poly-

urethane fire-retardant foam, foamed

in place. The bamboos are pinned and

and spot glued together at 50 cm intervals, with epoxy glue

and wooden dowels.

The framing model is shown below. The second storey floor

beams are all supported by interior partitions or columns, and

have clear spans of 3 meters or less - except in the family

room, where they span 4.50 meters between shear walls and impost

blocks, and are spaced close together to make up for the long

span. The roof beams span the full 4.80 meters between im-

post blocks, and are spaced at intervals ranging from 1.50 to

2.40 meters.

The second sto~ey floor is designed to carry 200 kg/m 2

2
(bamboo foam plank 15 kg/m , sulphur cement topping 45 kg/m 2 ,
2 2
second floor partitions 50 kg/m and live load 90 kg/m ). The
2 2
roof is designed to carry 80 kg/m (bamboo foam plank 15 kg/m ,
2 . 2
thin topping 20 kg/m and live load 45 kg/m ) . To put a third

-204-
 Second storey ceiling

First storey ceiling

storey on the house, additional beams will need to be inserted

(they can be slipped onto the impost block easily) , and the

topping on the roof increased.

At these loads, the beams have a deflection of less than

1/360 of the span, and can safely be plastered. (See table on

page 207 below) . Families who do not like the appearance of

the exposed bamboo can plaster them.

THE GENERAL PATTERN IS:

Context:

Short spans and light loads in countries where bamboo is

abundant and cheap.

Solution:

Beams may be made of bamboos (pinned and glued with epoxy)

to form a box which is filled with plastic foam . Spans may

-205-
range from 3 to 5 meters with corresponding variation in beam

spanning. Allowable loads are shown in the problem statement.

Problem:

Concrete beams are expensive, very heavy, hard to move

around, and hard to work. In many buildings, especially those
where people will be building for themselves (as in self-help
housing) beams need to be light weight, and easy to work. In
earthquake zones, it is also necessary to reduce dead loads as
far as possible. If bamboo is locally available and petroleum
resources allow local manufacture of urethane foams, then it
is possible to make lightweight bamboo/foam beams, with excel-
lent structural characteristics.

We have built three different beams of this type, and test-

ed them. It is clear from our tests that bamboo/foam beams of
this type are about as strong as softwood beams of the same
size. The most serious problem is deflection. Bamboo is ex-
tremely strong in tension, and the urethane foam makes the
beam section rigid; but the
bamboos tend to slip past each
other in horizontal shears .

In the third of the three

test beams , we pinned and spot
glued bamboos together with
epoxy glue and dowels . This
test beam was 20 cm wide , 40 cm
deep. We tested it over a
clear span of 3.50 meters. At
a uniformly distributed load
of 1300 kilograms the def lec-
tion reached 0 . 8 cm after an
hour, and showed no sign of
further creep 24 hours later.

We may use the formula:

3
Def lectionMax = (5/384)WL /EI

to obtain a value for EI, and extrapolate the following figures

for maximum allowable uniform loads, at various spans:

-206-
Clear span between supports Maximum allowable uniformly
(meters) distributed load, for beam
deflection less than L/360
(kg).

3.00 2200

3.50 1620

4.00 1240

4.50 980

5.00 800

(where the design criterion

is L/240, for unplastered con-
ditions, these loads can be
increased by 50%)

These beams will cost 100 soles per meter (compared with
about 200 soles per meter for comparable reinforced concrete
beams), and weigh about 20 kilograms per meter (compared with
50 kg per meter for a reinforced concrete beam of similar
strength). Furthermore, these beams can be cut with simple
tools: they can easily be lifted and installed by two men.

-207- ·
 It is important to note that the beam type described here
is by no means the last word in composite bamboo/foam beams.
Much development is needed to explore others which use differ-
ent indigenous materials in place of bamboo, other foams like
high-density sulphur foams, and new glues and bonding agents.•
The sketches illustrate some of these possibilities. For
discussion of urethane foam manufacture, see page 213.

-208-
COMPOSITE BAMBOO FOAM PL AN K

IN THE PROYECTO EXPERIMENTAL HOUSE, the second floor and roof

structure are bamboo/polyurethane foam sandwich planks laid

over beams. The outer skins of the plank sandwich are made

of 6 cm bamboos, and the core is two lbs.density polyurethane.

A sand-sulphur topping is poured after planks are in position

to form the upper walking surface, and the jointing between

planks (see page 214).

The planks are 15 cm thick (including the topping), 50 cm

wide and 5 meters long. They are supported by similarly con-

structed beams (see page 204) spaced at intervals between 1

I
meter and 2 . 40 meters, according to position in the structure. l;I

Since planks are 5 meters long , they act as continuous members

over at least two supports after the topping is poured over

them. If families do not like the exposed bamboo, the plank

will readily take plaster.

-209-
THE GENERAL PATTERN IS:

Context :

Short beam spacings and light loads in countries where

bamboo is abundant and cheap, compared to other materials .

Solution:

Floor and roof planks may be made from bamboo/poly-

urethane foam sandwich. Maximum span for this system is

approximately 2.50 meters unless panels have additional

thickness and reinforcing . Allowable loads are shown in the

problem statement .

·,: :;>::-:.:·i~gi;:: .: . .:~ --~:. / . ·:_~:. ~:~~?.~ .::. :_ _· :.·~:

g
~-------------- --·

Plank Section Plank Joint

Problem :

Conventional reinforced concrete beam and plank is expen-

sive and heavy . A number of recent experiments have shown that
sandwich planks with plywood , gypsum or cement asbestos skins
and polyurethane foam cores have enough strength to span 2-3
meters with normal live loads; they have been widely built and
tested in many parts of the United States . In a country where
bamboo is readily available, and wood expensive, it seems natural
to use bamboo as the outer skin of the sandwich instead of
plywood.

-210-
 We built a test ~lank, with half bamboos for the lower
skin, and 3 mm fiber board for the upper skin. This plank
performed very well in tests. At final failure the upper
skin failed, in shear; the bamboo held. The following table
shows the deflection test data.

These figures are for a center load, on a plank 70 cm wide,

over a span of 170 cm.

Load (kg) Deflection (cm)

45 .25
91 .50
136 .75
182 .95
227 1.15
272 1. 30
318 1. 50
364 1. 70
409 1. 90
454 2.10
546 2.55
636 3.00
729 3.50

This rudimentary plank, which has half bamboos in the lower

skin, and very little in the upper, is too weak. We recommend
a stronger plank, which has whole bamboos top and bottom.

-211-
 By means of the formula
3
Def lectionM ax (l/48)WL /EI

we may obtain a value of EI for the weaker plank. Reckoning

that the moment of inertia will be tripled in a plank with
whole bamboos top and bottom, we estimate that the stronger
plank will support the following loads, at the stated "spans:

Clear span between Maximum allowable uniformly

supports (meters) distributed load, for plank
2
deflection less than L/360 (kgs/m )

1. 00 2000

1. 50 590

2.00 250

2.50 128

3.00 74

These planks are e.xtremely light: they weigh about 1. 3

kilograms/m 2 , they cari be hand carried, and laid by two men.
Since they can easily be made in long lengths, it is advisable
to lay them over several supports, thus getting the benefit
of the negative mo ments. The urethane core gives them excel-
lent thermal and acoustic performance. The foam can also be
used as base for applying plaster or can be painted when de-
sired.

-212-
 Since the plank relies heavily on the use of polyurethane
foam, it is important to add a note on the manufacture of these
foams: particularly since the countries which are most likely
to benefit from the use of bamboo, like Peru, will have to
create urethane manufacturing capacity from scratch.
Capital equipment will cost $50,000 to $100,000. The
organization of the factory and one year's operation will cost
$200 000 - $250,000; with $100,000 of this amount going for
the initial inventory of raw materials. At these costs it will
be important to use polyurethane foams for other purposes too.
They can be used for beams (see page 204) e.g., for interior par-
titions, and in a slightly different chemical formulation, for
the manufacture of furniture, bedding and soft seating. For a
general discussion of urethane foams in building, see Structural
Potential of Foam Plastics for Housing in Underdeveloped Areas,
Architectural Research Laboratory, University of Michigan, Ann
Arbor, Michigan, November 1965.

-213-
SULPHUR REINFORCING AND TOPPING

IN THE PROYECTO EXPERIMENTAL, the connecting surfaces of blocks

which make up shear walls, bearing walls and columns are coated

with molten sulphur and fiberglass for tensile reinforcement.

Sulphur mixed with sand is placed 2.5 cm thick on the bamboo/

foam planks to create a walking surface, to create simple joints

between the planks and between planks and beams, and to give

continuity to the structure. Sulphur is also used as the water-

proofing agent in wet areas such as shower and toilet, sink

backsplashes, etc.

THE GENERAL PATTERN IS:

Context:

Low cost dry block construction and/or panel construction.

Solution:

Sulphur may be used as a reinforcing agent or bonding agent,

wherever tensile strength is required in block walls, shear walls,

-214-
planks and beams. It may be used by itself or with chopped

fibers, applied hot, or as part of a sulphur-sand grout.

Problem:

Jointing and reinforcing is a major part of the cost of

a block wall, particularly in a high earthquake zone. In 1969
in Peru, the cost of a mortarless block wall with sulphur joint-
ing is 20% less than the cost of a block wall with mortar and
reinforced concrete corners.

Test results show that the sulphur jointing has consider-

able tensile strength and that a sulphur jointed wall performs
at least as well as the standard block wall in low rise con-
struction. Data are presented in John M. Dale and Allen C.
Ludwig, "Sulphur Aggregate Concrete", Civil Engineering,
December 1967, pp. 66-68; in Allen C. Ludwig, Utilization of
Sulphur and Sulphur Ores as Construction Materials in Guatemala,
Southwest Research Institute, San Antonio, Texas, March 1969;
and in John M. Dale, "Sulphur-Fibre Coatings", The Sulphur
Institute Journal, Fall 1965.

Sulphur adds color and texture in those areas where it is

used. It may be tinted to produce colors other than yellow.

-215-
PLUMBING ACCUMULATOR

IN THE PROYECTO EXPERIMENTAL HOUSE, two accumulators above

the floor line provide traps, vents and waste disposal - one

for bathroom fixtures, and one for kitchen/laundry fixtures.

These accumulators are made of A.B.S. injection moulded plastic

and are designed to receive plastic piping, as used for all

plumbing pipes in the house.

THE GENERAL PATTERN IS:

Context:

Conventional water supply and waste system in any building.

Solution:

Accumulators are used to trap , vent and dispose of waste.

One accumulator serves as many as five fi x tures - the fixtures

must be grouped close together . Currently available accumulators

are made of A.B.S . injection moulded plastic , and are designed

to receive plastic pipes . (Patents are held by the Roth Russell

Corporation.)

Accumulator in place
Problem:

In traditional plumbing systems, a great deal of money,

labor and materials, goes into the construction of vents, traps
and clean out fixtures for waste disposal. Furthermore, once
they are installed, conventional waste disposal systems are
expensive to maintain and repair. An accumulator replaces all
the functions of conventional plumbing in one item with the
additional advantage that it allows above-floor installation
(good for slabs and future renovation) . Fixtures can be plugged
in easily, the accumulators may be used with left or right hand
fixture locations, and they may be relocated and reused. Accum-
ulators installed in Mexico and the United States have been in
use for at least four years.

-217-
CONTINUOUS ELECTRIC OUTLET

IN THE PROYECTO EXPERIMENTAL HOUSE, the electrical system is

a continuous elastomer strip , mounted onto the surface of the

walls .

THE GENERAL PATTERN IS :

Context :

Any house with normal electrical requirements.

Solution:

The electrical system is a continuous strip of highly con-

ductive elastomer, and can be mounted directly on wall surfaces .

.......,,.~ ~~~~~

'l ilt~ ~'t<f-/~

"""--4-.Jt

~~'/" ~\( thiaJ_

?'~~~~

-218-
It requires no special outlet boxes, but is a continuous female

connector: Fixtures and appliances can be plugged in at any

point. A special two pin plug connects appliances to the strip.

The system is called Power Distribution System, and has been

developed by International Technology, Inc .

Problem :

Traditional installation of electric wiring and outlets

is complicated and expensive, both in labor and materials. The
system described above is more convenient to use, and cheaper to
install. It allows appliances to be plugged in at any point
along the line. It is safer than the conventional system, since
it has a non-arcing plug to reduce the danger of electric shock
or fire, and the system is inherently water resistant. As far
as installation is concerened : There are no outlets, and no
in-wall wiring. The strip can be glued to any surface . It takes
less time to install than a conventional wiring system, and the
installer needs less technical competence to do it .

-219-