jason-users Mailing List for jason

Well...

I saw the you answer:

Jomi: Yes. I can remove the "true". I remember now why I use it: I did test
a if code and use the "true" to maintain it. Really it is not very
important.

Koster: You see my question about "false". I had need use the "false" to
cause the fault of plan. I will use the internal action now.

Thanks...


Em 28 de novembro de 2011 08:47, Jomi Hubner <jom...@gm...> escreveu:

> Hi Marcos,
>
> The simplest solution is to remove the true from the plan. I cannot see
> any reason to write "true" in the plan.
>
> Jomi
>
> On 15/11/2011, at 10:48, Marcos Côrtes wrote:
>
> > Hello again,
> >
> > I think I found a error.
> >
> > If I build a plan with a "true" sentence:
> >
> > +!g:true<-
> >     a;
> >     c;
> >     true;  //<---ERROR
> >     .
> >
> > the agent try to execute a action called "true".
> >
> > I do not know it an error. But I hoped that this plan ignore the
> sentence "true" because it mean the same a call for a internalaction or
> action with execution success. I too hoped if a use the "false" sentence,
> the system will fail the plan because the "false" mean the same of a
> internalAction/action fault.
> >
> > Thanks...
> >
> >
> > =================================
> > (in portuguese)
> > OLá novamente...
> >
> > Eu acho que achei um erro.
> >
> > Se eu construir um plano com a sentença "true" em su corpo...
> >
> > [mesmo plano anterior]
> >
> > o agente tenta executar a ação chamada "true" (e por consequência falha
> pois não epecifiquei a ação "true" no ambiente).
> >
> > Eu não sei se isto é um erro mas é um comportamento que eu não esperava.
> Eu esperava que o plano iria ignorar a sentença "true" pq isto significa o
> mesmo que uma internalaction ou action que executou e retornou sucesso. Eu
> também esperaria que se eu usa-se uma sentença "false", o sistema irá
> falhar pq falso é o mesmo que a falha de uma intenalaction/action.
> >
> > Realmente não sei se é isto que vcs queriam como comportamento ou se é
> um erro. Mas eu não espava isto.
> >
> >
> > Obrigado e espero ter ajudado..
> >
> > --
> > Ass: Marcos Côrtes
> > Mestrando Computação - UFF
> >
> ------------------------------------------------------------------------------
> > RSA(R) Conference 2012
> > Save $700 by Nov 18
> > Register now
> >
> http://p.sf.net/sfu/rsa-sfdev2dev1_______________________________________________
> > Jason-users mailing list
> > Jas...@li...
> > https://lists.sourceforge.net/lists/listinfo/jason-users
>
> --
> Jomi Fred Hubner
> Federal University of Santa Catarina
> Department of Automation and Systems Engineering
> PO Box 476, Florianópolis, SC
> 88040-900 Brazil
> http://www.das.ufsc.br/~jomi
>
>


-- 
Ass: Marcos Côrtes
Mestrando Computação - UFF

Thanks Andrew for the answer. Good points.
Although there is no point in adding "true" in a plan, as Jomi pointed 
out, Marcos is right, in my opinion. Since
+b : true <- true.
works (in fact +b:true. and +b. are syntactic sugar for that plan 
above), true should also work as an individual "no-operation" kind of 
formula. So in some sense it's a bug, but we don't promise to fix it. If 
you need just use .true and make it an internal action as Andrew suggested.

Cheers,

Rafael


On 28/11/2011 08:17, Andrew Koster wrote:
> Plans are a sequence of actions that an agent executes. Thus it is 
> unclear what you want "true" to do. I presume you mean it to do 
> something like "do nothing, but succeed". Jason has the .wait action 
> that could be used to do this (although I have not tried to run it 
> with .wait(0), maybe it must have at least 1 millisecond wait time). 
> Another option is to simply implement your own internal action and use 
> that.
>
> For your second question, false suffers from the saim problem, but in 
> this case there is a jason action that does exactly what you want: 
> .fail ends the current intention in failure.
>
> Cheers,
> Andrew
>
> El 15/11/11 13:48, Marcos Côrtes escribió:
>> Hello again,
>>
>> I think I found a error.
>>
>> If I build a plan with a "true" sentence:
>>
>> +!g:true<-
>>     a;
>>     c;
>>     true;  //<---ERROR
>>     .
>>
>> the agent try to execute a action called "true".
>>
>> I do not know it an error. But I hoped that this plan ignore the 
>> sentence "true" because it mean the same a call for a internalaction 
>> or action with execution success. I too hoped if a use the "false" 
>> sentence, the system will fail the plan because the "false" mean the 
>> same of a internalAction/action fault.
>>
>> Thanks...
>>
>>
>> =================================
>> (in portuguese)
>> OLá novamente...
>>
>> Eu acho que achei um erro.
>>
>> Se eu construir um plano com a sentença "true" em su corpo...
>>
>> [mesmo plano anterior]
>>
>> o agente tenta executar a ação chamada "true" (e por consequência 
>> falha pois não epecifiquei a ação "true" no ambiente).
>>
>> Eu não sei se isto é um erro mas é um comportamento que eu não 
>> esperava. Eu esperava que o plano iria ignorar a sentença "true" pq 
>> isto significa o mesmo que uma internalaction ou action que executou 
>> e retornou sucesso. Eu também esperaria que se eu usa-se uma sentença 
>> "false", o sistema irá falhar pq falso é o mesmo que a falha de uma 
>> intenalaction/action.
>>
>> Realmente não sei se é isto que vcs queriam como comportamento ou se 
>> é um erro. Mas eu não espava isto.
>>
>>
>> Obrigado e espero ter ajudado..
>>
>> -- 
>> Ass: Marcos Côrtes
>> Mestrando Computação - UFF
>>
>>
>> ------------------------------------------------------------------------------
>> RSA(R) Conference 2012
>> Save $700 by Nov 18
>> Register now
>> http://p.sf.net/sfu/rsa-sfdev2dev1
>>
>>
>> _______________________________________________
>> Jason-users mailing list
>> Jas...@li...
>> https://lists.sourceforge.net/lists/listinfo/jason-users
>
>
>
> ------------------------------------------------------------------------------
> All the data continuously generated in your IT infrastructure
> contains a definitive record of customers, application performance,
> security threats, fraudulent activity, and more. Splunk takes this
> data and makes sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-novd2d
>
>
> _______________________________________________
> Jason-users mailing list
> Jas...@li...
> https://lists.sourceforge.net/lists/listinfo/jason-users

Hi Marcos,

The simplest solution is to remove the true from the plan. I cannot see any reason to write "true" in the plan. 

Jomi

On 15/11/2011, at 10:48, Marcos Côrtes wrote:

> Hello again,
> 
> I think I found a error.
> 
> If I build a plan with a "true" sentence:
> 
> +!g:true<-
>     a;
>     c;
>     true;  //<---ERROR
>     .
> 
> the agent try to execute a action called "true".
> 
> I do not know it an error. But I hoped that this plan ignore the sentence "true" because it mean the same a call for a internalaction or action with execution success. I too hoped if a use the "false" sentence, the system will fail the plan because the "false" mean the same of a internalAction/action fault.
> 
> Thanks...
> 
> 
> =================================
> (in portuguese)
> OLá novamente...
> 
> Eu acho que achei um erro.
> 
> Se eu construir um plano com a sentença "true" em su corpo...
> 
> [mesmo plano anterior]
> 
> o agente tenta executar a ação chamada "true" (e por consequência falha pois não epecifiquei a ação "true" no ambiente).
> 
> Eu não sei se isto é um erro mas é um comportamento que eu não esperava. Eu esperava que o plano iria ignorar a sentença "true" pq isto significa o mesmo que uma internalaction ou action que executou e retornou sucesso. Eu também esperaria que se eu usa-se uma sentença "false", o sistema irá falhar pq falso é o mesmo que a falha de uma intenalaction/action.
> 
> Realmente não sei se é isto que vcs queriam como comportamento ou se é um erro. Mas eu não espava isto.
> 
> 
> Obrigado e espero ter ajudado..
> 
> -- 
> Ass: Marcos Côrtes
> Mestrando Computação - UFF
> ------------------------------------------------------------------------------
> RSA(R) Conference 2012
> Save $700 by Nov 18
> Register now
> http://p.sf.net/sfu/rsa-sfdev2dev1_______________________________________________
> Jason-users mailing list
> Jas...@li...
> https://lists.sourceforge.net/lists/listinfo/jason-users

-- 
Jomi Fred Hubner
Federal University of Santa Catarina
Department of Automation and Systems Engineering
PO Box 476, Florianópolis, SC
88040-900 Brazil
http://www.das.ufsc.br/~jomi

Plans are a sequence of actions that an agent executes. Thus it is 
unclear what you want "true" to do. I presume you mean it to do 
something like "do nothing, but succeed". Jason has the .wait action 
that could be used to do this (although I have not tried to run it with 
.wait(0), maybe it must have at least 1 millisecond wait time). Another 
option is to simply implement your own internal action and use that.

For your second question, false suffers from the saim problem, but in 
this case there is a jason action that does exactly what you want: .fail 
ends the current intention in failure.

Cheers,
Andrew

El 15/11/11 13:48, Marcos Côrtes escribió:
> Hello again,
>
> I think I found a error.
>
> If I build a plan with a "true" sentence:
>
> +!g:true<-
>     a;
>     c;
>     true;  //<---ERROR
>     .
>
> the agent try to execute a action called "true".
>
> I do not know it an error. But I hoped that this plan ignore the 
> sentence "true" because it mean the same a call for a internalaction 
> or action with execution success. I too hoped if a use the "false" 
> sentence, the system will fail the plan because the "false" mean the 
> same of a internalAction/action fault.
>
> Thanks...
>
>
> =================================
> (in portuguese)
> OLá novamente...
>
> Eu acho que achei um erro.
>
> Se eu construir um plano com a sentença "true" em su corpo...
>
> [mesmo plano anterior]
>
> o agente tenta executar a ação chamada "true" (e por consequência 
> falha pois não epecifiquei a ação "true" no ambiente).
>
> Eu não sei se isto é um erro mas é um comportamento que eu não 
> esperava. Eu esperava que o plano iria ignorar a sentença "true" pq 
> isto significa o mesmo que uma internalaction ou action que executou e 
> retornou sucesso. Eu também esperaria que se eu usa-se uma sentença 
> "false", o sistema irá falhar pq falso é o mesmo que a falha de uma 
> intenalaction/action.
>
> Realmente não sei se é isto que vcs queriam como comportamento ou se é 
> um erro. Mas eu não espava isto.
>
>
> Obrigado e espero ter ajudado..
>
> -- 
> Ass: Marcos Côrtes
> Mestrando Computação - UFF
>
>
> ------------------------------------------------------------------------------
> RSA(R) Conference 2012
> Save $700 by Nov 18
> Register now
> http://p.sf.net/sfu/rsa-sfdev2dev1
>
>
> _______________________________________________
> Jason-users mailing list
> Jas...@li...
> https://lists.sourceforge.net/lists/listinfo/jason-users

Hello again,

I think I found a error.

If I build a plan with a "true" sentence:

+!g:true<-
    a;
    c;
    true;  //<---ERROR
    .

the agent try to execute a action called "true".

I do not know it an error. But I hoped that this plan ignore the sentence
"true" because it mean the same a call for a internalaction or action with
execution success. I too hoped if a use the "false" sentence, the system
will fail the plan because the "false" mean the same of a
internalAction/action fault.

Thanks...


=================================
(in portuguese)
OLá novamente...

Eu acho que achei um erro.

Se eu construir um plano com a sentença "true" em su corpo...

[mesmo plano anterior]

o agente tenta executar a ação chamada "true" (e por consequência falha
pois não epecifiquei a ação "true" no ambiente).

Eu não sei se isto é um erro mas é um comportamento que eu não esperava. Eu
esperava que o plano iria ignorar a sentença "true" pq isto significa o
mesmo que uma internalaction ou action que executou e retornou sucesso. Eu
também esperaria que se eu usa-se uma sentença "false", o sistema irá
falhar pq falso é o mesmo que a falha de uma intenalaction/action.

Realmente não sei se é isto que vcs queriam como comportamento ou se é um
erro. Mas eu não espava isto.


Obrigado e espero ter ajudado..

-- 
Ass: Marcos Côrtes
Mestrando Computação - UFF

Hello jason developers.

I has a problem and maybe you can help me.

In short, I has built a change in the original BDI architecture and I add a
function F before the buf jason's function. This F function only can change
a subset of beliefbase and cannot change desire/intention base (for
example, it cannot call !goal).

My problem is that this functions are implements in java. But would be
interesting the user can build a jason plan with a especial tag.

For example: There is the plan:

@my_F_function[atomic]
+true:a & b & c
<-
    .internalActionA;
    action;
    -b_a; //"b_...." is the beliefs that this plan can change.
    +b_b;
    .

This plan, because use the tag "my_F_function" not will belong to the
Intention Base and each turn the architecture will check the beliefbase for
the a, b and c beliefs. If it is true, will run the plan (every atomic).


Could you help me talk some tips and when I begin read and change the
Jason? (ex. TS, methods, etc...)

Thanks,

I hope I'm not asking for much.

================================================
In portuguese:

Olá desenvolvedores,

Eu tenho um problema e talvez vocês poderiam me ajudar....

Eu desenvolvi uma extensão da arquitetura BDI conceitual e tenho de
implementá-la em Jason agora. Mas existe uma característica que eu não
consegui mapear totalmente:

Em resumo: eu adicionei um função "F" na arquitetura BDI que deve ser
executada antes da BUF do Jason. Esta função F só pode alterar um
subconjunto da BeliefBase e não pode mudar a Desire Base (não pode
adicionar uma meta com "!goal");

Eu consegui implementar a função F em java, mas seria melhor que eu pudesse
implementar um planos em Jason e pude-se "rotulá-los" para que minha
extensão do jason o considere como um plano que representará a função F (ou
vários planos).

[o mesmo exemplo de plano anterior]

Este plano, que usa a tag "my_F_function", não será adicionado na
IntentionBase. Além disto, o sistema sempre irá checar a BeliefBase se as
crenças a, b e c. Caso positivo, ele irá executar o plano de forma sempre
atomica.

Meu pedido é se vocês não poderiam me ajudar dando dicas de onde posso
começar a estudar e mudar o Jason (como por exemplo o TS, metodos, e etc...)

Obrigado e espero que não esteja pedindo muito, rs.


Ass: Marcos Côrtes
Mestrando Computação - UFF

Hi Lucas,

It isn't clear if you want to start up a whole Jason system from within 
a Java class, which is possible. Based on other experiences, I think you 
are talking about running Jason agents (programmed in Jason's variant of 
AgentSpeak) from within a Java system. It seems robocode doesn't have a 
server architecture and requires the code of each agent to be added to 
its own sourcecode before you compile. If that's the case, what you need 
to do is explained in Section 4.7 of the FAQ:
http://jason.sourceforge.net/faq
Unfortunately, that means you don't get an agent environment nor 
inter-agent communication already implemented for free. You'd have to 
implement those either from scratch or using relevant features of 
robocode. Let us know if that section of the FAQ is not what you are 
looking for.

HTH,

Rafael


On 04/11/2011 00:29, Lucas Del Castanhel Dias wrote:
> Dear Jason users and developpers
>
> I've been recently trying to integrate a Jason project with the 
> widespread robocode simulation tool (Available at 
> http://robocode.sourceforge.net/).
>
> However, I've been /experiencing/ a lot of trouble, specially because 
> this simulator has some restrictions with any kind of network 
> communication between a Jason project and the
> simulator.
>
>  I've tried to send information from a Jason project to a specific 
> group of robots using TCP or Java RMI, but both of the methods were 
> unsuccessful.
> We have also contacted the robocode staff to find a solution for this 
> issue but unfortunately it seems the simulator's architecture is 
> inadequate for such kind of approach.
>
> Therefore, we have decided to focus on Jason. I would like to know if 
> ,given a raw java class (A Foo.java), is it possible to invoke the 
> jason environment inside this class
> without the necessity of  creating a specific Jason project., i.e. 
> Invoke the jason environment in an average java project.
>
> Considering this idea is feasible, what should be done in order to 
> make it?
>
> I'm using the Eclipse Helios IDE and the latest version of Jason.
>
> Ps: If there's anyone who already had ever stumbled with the same 
> problem and got it solved from another approach and would like to 
> share I would be thankful.
>
> Yours sincerely,
>
> Lucas
>
>
> ------------------------------------------------------------------------------
> RSA(R) Conference 2012
> Save $700 by Nov 18
> Register now
> http://p.sf.net/sfu/rsa-sfdev2dev1
>
>
> _______________________________________________
> Jason-users mailing list
> Jas...@li...
> https://lists.sourceforge.net/lists/listinfo/jason-users

Hi Lucas, 

I think a first approach is to use only the jason reasoner for a single agent. Take a look at the example in the folder demos/using-only-jason-BDI-engine.

HTH,

Jomi

On 04/11/2011, at 00:29, Lucas Del Castanhel Dias wrote:

> Dear Jason users and developpers
> 
> I've been recently trying to integrate a Jason project with the widespread robocode simulation tool (Available at http://robocode.sourceforge.net/).
> 
> However, I've been experiencing a lot of trouble, specially because this simulator has some restrictions with any kind of network communication between a Jason project and the
> simulator.
> 
>  I've tried to send information from a Jason project to a specific group of robots using TCP or Java RMI, but both of the methods were unsuccessful. 
> We have also contacted the robocode staff to find a solution for this issue but unfortunately it seems the simulator's architecture is inadequate for such kind of approach.
> 
> Therefore, we have decided to focus on Jason. I would like to know if ,given a raw java class (A Foo.java), is it possible to invoke the jason environment inside this class 
> without the necessity of  creating a specific Jason project., i.e. Invoke the jason environment in an average java project.
> 
> Considering this idea is feasible, what should be done in order to make it?
> 
> I'm using the Eclipse Helios IDE and the latest version of Jason.
> 
> Ps: If there's anyone who already had ever stumbled with the same problem and got it solved from another approach and would like to share I would be thankful.
> 
> Yours sincerely,
> 
> Lucas
> ------------------------------------------------------------------------------
> RSA(R) Conference 2012
> Save $700 by Nov 18
> Register now
> http://p.sf.net/sfu/rsa-sfdev2dev1_______________________________________________
> Jason-users mailing list
> Jas...@li...
> https://lists.sourceforge.net/lists/listinfo/jason-users

-- 
Jomi Fred Hubner
Federal University of Santa Catarina
Department of Automation and Systems Engineering
PO Box 476, Florianópolis, SC
88040-900 Brazil
http://www.das.ufsc.br/~jomi

Dear Jason users and developpers

I've been recently trying to integrate a Jason project with the widespread
robocode simulation tool (Available at http://robocode.sourceforge.net/).

However, I've been *experiencing* a lot of trouble, specially because this
simulator has some restrictions with any kind of network communication
between a Jason project and the
simulator.

 I've tried to send information from a Jason project to a specific group of
robots using TCP or Java RMI, but both of the methods were unsuccessful.
We have also contacted the robocode staff to find a solution for this issue
but unfortunately it seems the simulator's architecture is inadequate for
such kind of approach.

Therefore, we have decided to focus on Jason. I would like to know if
,given a raw java class (A Foo.java), is it possible to invoke the jason
environment inside this class
without the necessity of  creating a specific Jason project., i.e. Invoke
the jason environment in an average java project.

Considering this idea is feasible, what should be done in order to make it?

I'm using the Eclipse Helios IDE and the latest version of Jason.

Ps: If there's anyone who already had ever stumbled with the same problem
and got it solved from another approach and would like to share I would be
thankful.

Yours sincerely,

Lucas

Hi Danilo,

thanks very much for reporting this problem!

Jason 1.3.6 changed the code of jade architecture and unfortunately the tutorial is not consistent. 

You have 4 alternatives to continue:

1. use the SVN version of Jason (where the bug and the tutorial were fixed).
	Instructions at http://jason.sourceforge.net/faq/faq.html#SECTION000101000000000000000

2. replace the internal action code to:
	   ....
           if (ts.getUserAgArch().getArchInfraTier() instanceof JasonBridgeArch) {
               // get a reference to the jade agent that represents this Jason agent
               JadeAgArch infra = ((JasonBridgeArch)ts.getUserAgArch().getArchInfraTier()).getJadeAg();
	    ....

  and use the jason.jar available at
	http://www.das.ufsc.br/~jomi/x/jason.jar

3. use jason 1.3.5

4. wait for jason 1.3.7 :-)

All the best,

Jomi


On 31/10/2011, at 17:19, Danilo Pianini wrote:

> Hello,
> I tried to follow the document at http://jason.sourceforge.net/mini-tutorial/jason-jade/ using Jason 1.3.6, but it seems to be no longer consistent. It complains that it can only run on a Jade platform, and uncommenting the instanceof control throws the following:
> 
> Error in internal action 'jadedf.register'! java.lang.ClassCastException: jason.infra.jade.JadeAgArch$JasonBridgeArch cannot be cast to jason.infra.jade.JadeAgArch
> 
> The inner class JasonBridgeArch, having no public modifier, is not accessible from an external Java code. Moreover, all this is not consistent with the API description:
> http://jason.sourceforge.net/api/jason/architecture/AgArchInfraTier.html
> JadeAgArch is actually a subclass of JADE's Agent (http://jade.tilab.com/doc/api/jade/core/Agent.html), which obviously does not implement the interface AgArchInfraTier.
> 
> Is there something wrong or it's all my fault?
> 
> Cheers!
> 
> -- 
> Ing. Dott. Danilo Pianini
> Contract Researcher
> 
> Contacts available at:
> http://www.unibo.it/docenti/danilo.pianini
> http://apice.unibo.it/xwiki/bin/view/DaniloPianini/

-- 
Jomi Fred Hubner
Federal University of Santa Catarina
Department of Automation and Systems Engineering
PO Box 476, Florianópolis, SC
88040-900 Brazil
http://www.das.ufsc.br/~jomi