11-12 J une 2009, NAi, Rotterdam.

Hybrid Architectural Archives

Davi d Peycer
Centre darchives darchitecture du XX
e
sicle
Cit de larchitecture et du patrimoine
127, rue de Tolbiac, 75013 Paris
00331 4585 1200, dpeycere@citechaillot.fr

THE ARCHITECTURAL PRACTICES
AS FIRST CURATORS OF THEIR ARCHIVES
Introduction
During this conference on hybrid architectural archives, I will be using the term hybrid in a very
particular way. Many of the talks we have heard over the last two days have been speaking about
computerised architectural archiving projects which are hybrid in the sense that they partly require use
of archives, or in the sense that they can be perceived as archives (i.e. sources of information) whilst
not representing the organic output of one persons actions. I want to take a more traditionalist
approach and talk about architectural archives in their most fundamental sense in the sense that an
archivist would refer to them in other words records produced and received during the exercise of a
particular function. In this sense, if we refer to hybrid archives, their hybrid nature refers to the fact
that they contain both conventional records and digital records. And what we are interested in here is
the digital aspect of these records.
I want to talk about production and management by the producer of computer-generated archives from
architectural practices. This study is based on two surveys carried out in 2003 and 2004, which I will
talk about shortly, and on a number of in-depth visits to Paris architectural firms in 2008 (Architecture
Studio, Paul Chemetov AUA, Renzo Piano Building Workshop, and J ean Nouvel Architects).
Here are the basic points I would like to stress. Although some conclusions are drawn from French
examples only, I am assuming (and this may be an interesting point to discuss), that they are more
broadly valid.
Architectural practices not only produce records, but they are their own curators. They are obliged to
store some of them for decades, which is much longer than what a public agency does with its
records; they have to dispose of them, either under the pressure of urgency, or along carefully thought
out lines; and for the electronic records, they have to modify their physical packaging and to migrate
them. Every practice is its own master and has to find its own way out of the sea of issues raised by its
records.
It should be stressed that, from the point of view of professional archivists in specialised archive
repositories, the practices are already doing a part of our job, taking decisions on the fate of each
document or file. We consider that they are fully responsible for the records management step, but
also, looking further forward, a part of the long-term archives management. Although not all of their
decisions result from a balanced reflection, some do, and both types of decision are of interest to us.
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 2
Concerning digital files, while the philosophy (if any) differs significantly from practice to practice, the
results might not be so vastly different. Two issues seem to be particularly relevant for us: should the
master copy the one to be kept the longest be digital or paper-based, and what format policy can
best provide access to the files in the future. Besides, and outside of the scope of this paper, how far
ahead is the future?
Today we observe the tendency which, I believe, is relatively new to use a wide (although perhaps
narrowing) variety of programs to create the files, but simultaneously to create at a given moment in
time, a digital output of each original, in a very small series of formats. Then, often, in addition, there is
also a paper output: even large studios with excellent procedures on file production, migration and
storage seem to trust paper above all and put their care on systematically printing the files and storing
the prints. The creation of these outputs (digital and/or hard copies) usually happens at a critical point
of the project typically, the completion of a project phase , and happens only to files retained for
preservation. Thus, the first and sometimes the only choice is made at a very early stage of the
records life, even before the architectural project in itself is completed.
Even if every practice has to define its own rules about managing its records, these rules, if applied
consistently within each practice, should help us archivists a lot in our task of accepting and managing
an archive.
Before I get properly started, I want to stress that this presentation is based on six years of work by
the Gau:di programme Architectural Archives working group which I helped to coordinate. I would like
to thank the whole group, which I am representing at this conference. Gau:di is a European
programme whose aim is to bring together the experience of architectural archiving institutions in
different European countries. Over the two three-year cycles of this programme, from 2001 to 2008,
specific work was carried out regarding architectural archives, drawing together a number of
representatives of specialised archive centres. The NAi and IUAV from Venice, which are represented
here today, and the Mendrisio Architectural School from Switzerland, the RIBA archives in London, the
architectural museums of Finland and Norway and the Deutsches Architekturmuseum in Frankfurt
were all part of the programme. Some of the work carried out by this group has been published on the
website http://archivesarchitecture.gaudi-programme.eu/. One of the groups activities was to discuss
and share experiences with respect to electronic architectural archives and the way that these records
are managed within archive centres. We quickly realised that we were absolute beginners in this field.
When the programme started in 2001, all we knew was that we could not keep on ignoring this aspect
of archives for much longer. In 2007, we organised an international conference on the subject and the
proceedings of this conference were published in autumn 2008.
I would also like to emphasise how indebted we are towards the Art Institute of Chicago and the
unrivalled study that it carried out over the same period. Following extremely in-depth analysis, the
institute offers on its website an exhaustive list of file formats for migrating the native formats used by
architects along with recommendations regarding records management within architectural practices.
I want to start by talking about the production of records in architectural practices from an archivists
perspective and then go on to talk about the way in which firms managed their records today.
1. The production of digital records in architectural practices
The digital archives produced by architects will depend significantly on the way their production
structures operate. Some of the key features are as follows:
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 3
Record-producers are private-sector structures of variable size, but often very small. There are
very many such producers, who often work with a fairly unsophisticated approach. In France, very
few architectural practices have more than 150 employees. Such firms have a long history of
mergers, associations and temporary partnerships, particularly in working towards architectural
competitions.
The records produced often feature complex structures. They include graphic files in two or three
dimensions, structured as layers, with associated databases. They require other documents in
order to be complete. A large number of software packages are used to generate such records,
hence the broad variety of outputs (drawings, written documents, tables, videos, sound files etc.).
The vast majority of architectural practices started using computer technology over a very short period
of time at the end of the 1980s and start of the 1990s. Two surveys in 2003 gave an insight into the
way in which this IT take-up and change of medium took place. The Europe-wide survey found that
one-quarter of agencies surveyed said they became computerised between 1980 and 1989. 1994
seems to have been the last year in which a practice could still be set up without immediately requiring
computer hardware. As we will see, when we refer to computerisation, it is often the practice staff
who gained access to computer equipment, but the founding practice partner did not necessarily start
using this technology.
The relationship between architects and IT, between architecture and cybernetics, is much older and
goes back many decades earlier in the twentieth century in tangible, experimental, imaginary or
fantastical ways. The backdrop to this relationship is the idea which has still not become a reality
of one day no longer requiring paper, achieving the Paperless Studio which, by the way, was the
name of a programme of study at the Columbia University, New York. Do any paperless practices
exist today? This is a twofold question. From an archiving perspective, is it possible to only store
computer files and no physical records? From a creative point of view, is it possible to carry out all
stages of design on screen and using computers only? Although it may be true that nothing is
impossible, it should nevertheless be noted that despite the ubiquity of information technology,
computers remain a tool that has not actually brought about the metamorphosis in architectural
practice and design that the utopic thinkers dreamt of in the years between 1940 and 1960. What is
today referred to as non standard architecture is a borderline case and may be an exception to what
I have just said I will come back to this shortly. The use of a computer by the practice partner or the
project managers i.e. the creative architect may be the last bastion of resistance. Whatever the
architects generation and whether or not they were trained in computer technology during their
studies, some designers still work only by hand and pass on their sketches to their teams to
implement. For drawing, in addition to the conventional office software used in any company and e-
mail, which is used in the same way as everywhere, architectural practices have gradually taken up
computer-aided design tools (CAD), 3D modelling tools, image editing and animation (video) solutions.
A number of software packages or formats are now so widely used that they have become the de
facto standards (e.g. AutoCAD, with its DWG format, which became a near-standard solution by
around 2000) or official standards (e.g. the PDF format from Adobe). These standards facilitate
transfer operations for the files that are generated today. However, the first ten or fifteen years of
computerised architectural output in many practices may well become inaccessible in the near future
and in some cases this is already so. In 2003 however, three-quarters of European practices surveyed
claimed they were able to access the computer files created in their firms 10 to 15 years earlier.
In the area of design, the situation is a lot less standardised. In 1996, a survey looked into the use of
IT in architecture outside of computer-assisted drafting tools, looking both at the production of
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 4
presentation documents and initial project design. Responses given at the time highlighted the magic
and fascination of images available in image banks or pictures that could be created prior to a project.
However at the same time there was a lot of mistrust with regard to the ease of this process, which
could quickly get out of control.
Less than ten years later, exhibitions such as Architectures non standard at the Centre Pompidou in
Paris (2003-2004), demonstrated that design tools, as used by a small minority of architects, are
capable of changing not only the form, as had been anticipated as early as 1996, but also the build
and manufacture methods used in architecture, by developing non-standard modules (with or without
the use of computer-numerically-controlled manufacturing or computational processes). These
exhibitions often highlighted theoretical research projects, which used 3D modelling and animation
software from the domains outside of conventional architectural work in a new and experimental way.
Beyond such experimental projects, it is also important to highlight the continuing importance of
physical documents sketches, design models etc. in architectural design.
Digital architectural records thus feature a large quantity of drawings, graphic representations with
dimensions in other words CAD files. However, programs such as Photoshop, Excel, Word and
PowerPoint clearly have a major importance in architectural practices (as highlighted both by the
European survey and the Chicago study) to the extent that the architects surveyed mentioned them
even more often than AutoCAD, the leading CAD package. Architectural design is a process that
therefore cannot be reduced simply to the use of CAD files.
However, these CAD files are the biggest issue in terms of long-term preservation, raising issues of
transfer and migration transfer from one medium to another, migration to more up-to-date, optimised
formats or formats specifically designed for preservation purposes. In the case both of transfer and
migration, not all features of the original file can be preserved.
2. Records management within architectural practices
It took architectural firms a number of years to realise the need to organise their IT filing system in a
much stricter way than their paper records had been organised. Internal rules needed to be drafted. In
France such rules started to be put in place in the early part of this decade, in simple or complicated
forms, and such rules are obviously implemented to a greater or lesser extent by practice staff. To
give an example, the Renzo Piano practice, an international architectural firm with more than one
hundred employees (which means it is a big firm for France) only drew up such rules in 2007 and has
been gradually implementing them over the 2008-2009 period, with a major concern to avoid cutting
across the individual approaches of their staff members. When shared data storage systems are used
(such as external servers and drawing archives), these systems help impose a structure. This
organisational approach may, for instance, be embodied in a tree structure for organising files within a
project and which can be implemented for all projects, naming rules, both for files project code,
phase code, drawing type code applicable both to files and the layers within the files.
In addition, the development of a quality management system and ISO 9000 certification to guarantee
its implementation, are a real help to architectural practices in structuring their data and creating a
solid, reassuring archive structure. The ISO 9000 standard can be used as an inspiration without
being fully implemented and without certification necessarily being the final goal.
These days, architectural firms tend to distinguish two types of digital records and formats:
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 5
source, native, exchange files (for exchange inside the firm and with project partners), which
are used in the short-term,
and output, distribution, print or archiving files.
Internal guidelines recommend that the second type of file be systematically generated from the first
type and that clearly differentiated objectives, statuses, production times and life cycles be specified
for the different types of files.
Let me first of all talk about the source or native files.
Vector-type drawing files accumulate within a single file the results of an ongoing differential process
of information and knowledge accretion, whereby this single file ends up constituting a varied
collection of expressions of reality, knowledge and memory"
1
. I Three features of such files make
preservation and migration rather difficult their layers, cross-referenced files and pen files. Layers
are used to design a variety of different embodiments (outputs or print-outs) for a single file. In
AutoCAD, until the Presentation tab was developed in the LT 2002 version, the architects preferred
view could not be reconstituted. Cross-references (xref) may lead to blank spaces in the output file if
they are not preserved with the file in the original file structure. Finally, pen files were only included
with the drawing file from version LT 2002. Previously, a user needed to know the codes used by the
architectural firm in order to reconstitute the initial appearance of any drawing. In France, for example,
there is no generally accepted convention for these codes nor is there necessarily a standardised
usage within a single architectural practice individual project managers or even individual draftsmen
may use such codes differently. Obviously these days, individual habits are starting to give way to
common standards, which are described in documents referred for example as a firms graphic
charter. However, such habits do, surprisingly, live on, as they can still be observed in some major
firms to this day.
Output or archiving file formats
Drawing files are used in two ways in the project development process. Firstly, they are used to draft
the plans for which they are being designed at any given phase of the project (e.g. preliminary design
phase), and to make any edits to this drawing. Secondly, it may be used as a starting point for the
drawing of the same item in the next phase, in which this drawing will be issued a greater level of
detail (e.g. detailed design phase). Once the project has reached a subsequent phase, the drawing file
is no longer used as a production tool and the practice only requires printable versions of the drawing
for each phase.
This has led architectural practices to implement the following methodology, which is often described
in internal guidelines. At the end of each project phase, the project manager sends off or saves in a
particular location on the server, all plans for this completed phase in a preservation format from which
the files can be directly printed out. These are "frozen" formats in which the drawing cannot be edited.
They may, at this stage, be defined as the only official records within the practice, as opposed to the
source files which can still be edited. In archiving terms, these records comply with the criteria of
integrity and authenticity, which are fundamental for long-term preservation. It is therefore

1
Stphane Loret, Pascal Garret, Crvilles, un systme darchivage, gestion et diffusion dimages sur la ville ,
Architecture et images numriques, 2008, p. 509.
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 6
recommended that AutoCAD files be converted into DWF format, Word or Excel files into PDF format
and Photoshop images into PDF or J PG formats.
DWF seems to be coming to the fore as the standard distribution format for vector drawing files
particularly from AutoDESK applications. This will continue to be the case unless it is overtaken by a
version of the PDF format. DWF is generated from within AutoCAD by using the Publish function.
The PDF format has become a standard and is used as an exchange format between different Adobe
software packages (Photoshop, Illustrator, InDesign, etc.). It can be generated from AutoCAD
including in batch mode using software such as Adobe Writer 8 and PDF Creator.
Other CAD file conversion formats exist, mainly for the process of file exchange. In particular, the DXF
format should be mentioned, an AutoDESK proprietary format which is generic and can be used to
exchange files between all CAD programmes. The IGES, STEP, IFC, STL and the new PDF/E formats
are used for 3D files.
In some architectural firms, the guidelines state that the DWG files must be stored in parallel, each
one accompanied with a ZIP file containing all the pen and xref files.
Many architectural practices also have PowerPoint files (PPT) which include files of many formats
(CAD files, pictures, animations etc.). Some firms systematically use PPT files to preserve an
overview of their projects in the same way that in the past key project documents were kept in folders
or bound portfolios. At the Renzo Piano practice, each project manager is responsible for defining one
or more iconic representations at each stage. This may be a plan view, a section or any other
representation.
E-mails raise specific issues which are currently not well dealt with. Recommendations are often made
to print them and attach them with the paper file for each project. Digital conversion and presentation
is not always required.
Various firms store one or more file types in parallel, either chiefly in native formats or chiefly in output
formats or both. Sometimes part of the project files are gathered together in PPT files.
Some practices highlight the need to select files in order to avoid keeping all files generated during
project development. The total file volume within architectural firms runs into terabytes, leading some
people to pay attention to the amount of data stored. One rule that is frequently mentioned, but which
does not appear to be consistently implemented, is that parallel versions of the same drawing should
never be created, but the same name should always be used in order to overwrite each successive
iteration. Even with such rules, it is often recommended that each person involved in the project, and
the project managers, regularly delete everything that is not fundamental (at the end of each phase).
The key moment for data management actions seems to be the end of each phase whenever any
deliberate archiving activity is implemented. It would however seem that conventional project
management phases, as found in archive records, have significantly changed over the last twenty
years, with the use of IT tools. The successive preliminary sketching, design and construction drawing
phases have been blended into a continuum, since digital data can easily be reused and imported
from one phase to the next, even going as far as model making and the information used in building,
since the advent of computer-numerically-controlled (CNC) manufacturing processes.
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 7
Storage media and metadata
Most practices systemically store their data on the hard disks of a server, with back-up and security
procedures that these days should be well documented and monitored. Files are also often duplicated
onto other digital storage media (CDs or DVDs). However, firms seem not to have specific, well-
monitored rules for file organisation prior to creation of the disk (often at the project managers request
at the end of each phase), or procedures for actually burning the disk. Likewise, CDs are rarely
subject to periodical checks.
Other types of storage media, such as cassettes, tapes, cartridges and floppy disks, seem to have
been phased out. Some practices have old storage media of this type and have still not organised
transfer processes. They run the risk of losing the files stored in this way. However, this situation is
never viewed as a major issue.
The use of metadata other than the filename and sometimes the date of creation also seems to be a
rare practice, although at times it is referred to in the guidelines.
What is the original? Should hardcopies be kept?
Many digital files are printed out or otherwise output as a hardcopy. In 2003 the question of What is
the original, the hardcopy or the electronic version? seemed to be a relevant question for architects
and the majority of them stated that the hardcopy was considered to be the original. At the time, we
expected rapid changes in this perception.
The real question is whether this shift has now taken place? What is striking these days, in France at
any rate, is that even those practices who place the most trust in their hardware and IT procedures
consider that the hardcopy is still the chief record to be preserved or archived for the long-term
(sketches, physical models and print-outs of CAD drawings). The storage of such items is organised
as carefully as their digital records. What is less clear is whether such hard copies are output using
materials (paper, ink etc.) that are suitable for long-term preservation. The printed records that I
observed in France are stored in the open-air with significant exposure.
Training architects in the practices
Meetings between archivists and architects have raised a number of questions about methodology.
Since electronic records management begins at the time of creation, it seems vital for all document
producers i.e. almost every employee in an architectural practice to be aware of a number of
general principles and specific rules that have to be systematically implemented. It would also be
useful to raise awareness of the continuum between the document created today and the file that will
be stored in the firms archives as evidence or as a historical record in a somewhat more distant
future.
More often than not, architects display fairly limited interest in the long-term preservation of their
archives. Forming a historical archive which could be transferred to an archiving institution is rarely
talked about as a goal and sometimes explicitly rejected. Having said that, firms always find a way to
document their projects. Presentation documents are generated throughout every project. These
records are intended for a wide audience and are often managed by the person responsible for press
relations. The architects often trust these documents, which are available in commonly-used but
non-editable formats such as PDF and J PG, to provide any record that will be used for future
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 8
reference. The contractual and administrative documents are rarely needed beyond the legal
guarantee period, which is ten years in France.
In-house management of archives should draw on the principals of records management as laid down
by the ISO 15489 standard. Architectural agencies and any records-generating organisation can
use this standard to define, internally circulate and implement specific rules in the following areas that
have been listed by the AIC:
File naming, in particular codification of project names, phase names and graphic file types (plans,
sections, elevations etc.).
Print-out conventions (pen codes).
Timetables for the validation and selection of documents to be disposed of (and the person
responsible e.g. project manager), with a report given to an external manager (at the end of
each phase, ideally).
Arrangements for CD or DVD transfer, where applicable.
Monitoring of the records stored (e.g. opening the CDs on a periodical basis).
Lists of rules to be implemented by architectural practices were drawn up following the two
abovementioned surveys in 2003, and can be downloaded from the Gau:di working groups
documents and from the AICs records. The findings of the latter organisation are summarised in the
proceedings of the 2007 conference.
Conclusion
My observation is that between the initial surveys in around 2003, approximately ten years after the
mass roll-out of computer technology, and my most recent visits, the issues raised by electronic
records and archiving have somewhat diminished in architectural firms. Although rules for production
and storage do not exist everywhere, there are now guidelines for drafting and implementing such
rules and the practices are aware of their need to make an effort in this area. These days, unless there
is an exceptional occurrence, any losses likely to occur will not prevent archivists and researchers
gaining an overview of and an insight into the output of each firm. This finding is clearly positive in
general terms. However if one looks closer, a number of issues remain. A second issue has now been
raised although architectural practices are becoming credible curators of their own records, the
methodologies for transferring their records to long-term archives remain to be defined. The 2007
conference provided an opportunity to map the problems, current research and contradictions in this
area. The solutions presented were mostly applicable to small volumes of records or else were clearly
experimental projects. For the moment, we still have no generic long-term solution that is ready to be
implemented.

Bibliography
This paper is based on several presentations at the Architecture and Born-Digital Archives
conference, Paris, 8-10 November 2007, the proceedings of which have been released as:
Architecture et archives numriques. Larchitecture lre numrique : un enjeu de mmoire /
Architecture and Digital Archives. Architecture in the Digital Age: a Question of Memory (bilingual,
Hybrid Architectural Archives, Rotterdam, J une 2009 David Peycer 9
edited by David Peycer and Florence Wierre, coordinator Carole Koch), Gollion/Paris, InFolio/Cit de
larchitecture et du patrimoine/Institut national dhistoire de lart, 2008
The 2003-2004 Gaudi survey: http://www.architecturearchives.gaudi-
programme.eu/fichiers/t_pdf/5/pdf_fichier_fr_GAUDI_Report.pdf
The 2003-2004 Art Institute of Chicago survey:
http://www.artic.edu/aic/depts/architecture/dddreport/0B_START.pdf
ICA Committee on Current Records in an Electronic Environment, Electronic Records: A Workbook for
Archivists, April 2005 (coll. Studies/tudes n 16), Chapter 6.
David Peycer, Les archives numriques des agences darchitecture , Comma, International
Council on Archives, n xxx, 2009 (forthcoming: a detailed version of this paper, in French).