Ciao-users April 2002

ciao-users@cliplab.org

5 participants
7 discussions

Re-Re: Ciao-Prolog and Agents
by Jane Loans 03 May '02

03 May '02

Thanks by the help. But, it gives me the impression of the fact that the active modules are not active, but "passive", since: - Therefore I have proven, an active module can't have a predicate "main/0" or "main/1" (and if it has, not serves to him as nothing because upon starting the active module, the the "main" predicate is not executed). - All the agents (active modules) on started are listening on its port to the arrival of some message. - Because the two previous reasons, I must to have a single agent that make the paper of "initiador", the one which can have defined the predicate "main". This initiator passes a message to the "agentI", later on the "agentI" calls to the "agentJ", and so on. What mean with this, is that IN ALL MOMENT CAN HAVE 1 COMMUNICATION (AMONG 2 AGENTS); IT'S IMPOSSIBLE THAT, FOR EXAMPLE, THEY HAPPEN THESE 3 COMMUNICATIONS SIMULTANEOUSLY: agent1-------->agent5 agent3<--------agent12 agent9<--------agent7 I am in what is certain? Something else: Could I use the Module Qualification instead of the Higher Order? That is to say, to use send(I, Message, Receiver) :- Receiver:receive(Message, I). instead of send(I, Message, Receiver) :- Receiver(Message, I). Finally, you said me that you probed the code that I sent (with you modifications). Could you give me the source code (the ".pl" files) that you have proven? __________________________________________________ Do You Yahoo!? Yahoo! Games - play chess, backgammon, pool and more http://games.yahoo.com/

2 1

Re: Delete numbers of a atom of four numbers
by Manuel Carro 24 Apr '02

24 Apr '02

>> If I have a atom of four numbers, for example, '2002', How I can delete >> the first and the second numbers, for obtain, in the previous example, >> '02'?. I have a database of bibliographic records, compound of facts, >> with the date in format dd/mm/yy, and I need compare the system year >> with the year of the records in database. But the predicates get_time/2 >> and convert_time/8, provide the year of the sistem date with four >> numbers. For this, I want convert the year format in four numbers, to >> format in two numbers (the two last numbers of the system year). ?- atom_codes('2002', [_,_|_L]), atom_codes(Result, _L). Result = '02' ? In any case, a bibliographical database with only two digits for the year does seem to make much sense. Books have been published for centuries. MCL ________________________________________________________________ [...] put Windows back into its place as an overpriced Nintendo.

2 1

Re: Ciao-Prolog and Agents
by Francisco Bueno Carrillo 22 Apr '02

22 Apr '02

Dear Jane, It is quite easy to turn Ciao active modules into agents for the kind of application you mention. Active modules are run independently and communicate through messages. Of course, execution is non deterministic, too. Ciao also runs in several platforms (including Win32 and Unix), so you will find your application easily portable. The Prolog code you send as receipt guard is basically all you need. Here is a slightly modified version that we have been running here without problem: case(shutdown, _) :- !, halt. case(add(NumList), Sender) :- sum_list(NumList, Sum), send(Sender, sum(Sum)), !. case(_, _). We have taken out message give_me_your_id and variable I, since we have assumed that the agent Id and I are the same thing, and we have used the module name as such an Id, which is convenient when using active modules. Thus, you identify an agent by the name of the corresponding active module. This is done transparently by the module "send" below, which implements the send/2 predicate. We have also assumed that any agent might want to call any other, and that the number of agents is known in advance. Given this, it is convenient to use an auxiliary module that will take care of agent communication. It will look like: :- module(send,[send/2],[actmods,hiord]). :- use_module(library('actmods/webbased_locate')). :- use_active_module(agent1, [agent1/2]). :- use_active_module(agent2, [agent2/2]). ... :- use_active_module(agentN, [agentN/2]). :- meta_predicate send(?,addmodule). send(Sender, Message, I) :- Sender(I, Message). Basically, you need this module to declare the agents in your application: agent1, agent2, ... We have used as communication protocol the "webbased" one, but you could use any of the others. With "webbased" you will need to have a process acting as name server. Given the above, your agents will look like: :- module(agent1,[agent1/2]). :- use_module(send). agent1(Sender, Msg) :- case(Msg, Sender). ... code for case/2 ... and all you have to do is to compile them as active modules: ciaoc -a actmods/webbased_publish agent1 We have been able to run such an application at our site. Your problems might be caused by the name server (if you're using the Ciao distribution right away, the name server is supposed to be run at our site, and we are not currently running any). You will need to set up an appropriate name server at your site. We can help you with this if you have problems (you are probably right that the documentation is not very clear in this point). It is not clear to us from what you say: >> and when the agent end of executing the instruction >> underway, agent attends the arrival of the message >> (is as an asynchronous event that triggers the >> attention of the agent). whether you want agents to attend several messages at the time or not. It is possible to do it using threads. With the above code, however, agents process only one message at a time (and they are enqueued waiting for the agent to be idle). Paco Bueno & the Ciao team

1 0

Stop a query or a predicate
by David STRAINCHAMPS 19 Apr '02

19 Apr '02

Hi, I just want to know if it is possible to stop a query automatically after About 15 seconds when the predicate in cause by the query don't find a solution or a no response. I don't know if I explain well my problem Thanks for answers David

1 0

Update on web site
by MCL 11 Apr '02

11 Apr '02

Dear Ciao Prolog users, we have became aware of a quite serious bug which affected the performance of the Ciao Prolog system. The most relevant effect was compiling times much longer than actually needed. A new snapshot correcting this bug (among others) is available on the web site: htpp://clip.dia.fi.upm.es/Software/Ciao Best regards, MCL & the Ciao Prolog team -- +--------------------------------------------------------------------------+ |Manuel Carro --- C.S. School --- Technical University of Madrid (UPM)| | Email: mcarro(a)fi.upm.es --- Ph: +34-91336-7455 --- FAX: +34-91336-7412 | |http://lml.ls.fi.upm.es/~boris --- http://clip.dia.fi.upm.es/Software/Ciao|

1 0

Re: setrand sisctus predicate ?
by Manuel Carro 03 Apr '02

03 Apr '02

Should be pretty easy: something similar to setrand(rand(X, Y, Z)):- Seed is X * Y * Z, srand(Seed). There is no guarantee that you'll get the same sequence, though. MCL ______________________________________________________________________________ Miss a couple of car payments and you'll see if someone cares if you're alive.

1 0

setrand sisctus predicate ?
by idees folles 03 Apr '02

03 Apr '02

ciao prolog don't have the sisctus predicate setrand(rand(X,Y,Z). i have a programm with it how can i do ?

1 0

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

Ciao-users April 2002