Hi,

As of now, I don't have any knowledge on standards used in the agents 
world ... I also think I don't have the time to tackle that on short 
notice. But I've some ideas on an agent framework that's written in 
Perl,e.g. itinerary, transport and the agent object by itself. Although 
you should separate the agent from the transport over which it has been 
transported, I would like to experiment with jabber as a possible 
transport medium ...

Maybe I'lll brew something myself, but I'm afraid it won't be standards 
compliant ...


Johan

Steve Purkis wrote:
> On Fri, 20 Dec 2002, Paul Barry wrote:
> 
> 
>>johanvdb wrote:
>> > It seems that the Perl Agent framework dates from 1998. Are there any
>> > new initiatives under way? Does somebody wants to take up the
>> > development of the framework?
>>
>>I documented the current state of Agent Perl in chapter 6 of my book
>>(link below).  I did have a student work on some enhancements last year,
>>and he was supposed to send his code to Steve Purkis.  I'm not sure if
>>he did this, as his efforts were quite disappointing.  Steve had
>>indicated that he would like to get the ball rolling again, but also
>>stated that he wasn't in a position to dedicate a lot of time to it.
> 
> 
> I didn't receive any updates, unfortunately.  But it's true - I'd like to
> see the ball rolling and would be willing to help, though my time is
> limited ATM.
> 
> Anyways, there has been talk of new initiatives, but as of yet, none are
> under way...
> 
> I've had some interesting chats with Ben Evans sporadically over the past
> few months where we'd talked about a new framework, in particular trust
> mechanisms, and Perl modules for OWL & DAML/OIL, which in turn would
> require a namespace-aware RDF framework...
> 
> I'd also had some conversations a while back with Matt Lawrence about
> modelling a new framework on the work done by FIPA (www.fipa.org - which
> contains more ideas than you could shake a stick at... :), and with Tim
> Finin on the evolution of KQML/KIF to OWL. (All of this was sparked by
> Paul's interest, so thanks Paul!)
> 
> If a new initiative is going to happen, I'd suggest it start by:
> 
> 1. Reviewing the current system, picking out good & bad points.
> 2. Talking about what's going on in the world of agents, and
>    making proposals about what a perl agent system should be/do.
> 3. Decide & agree on what to build.
> 4. Build it. (the easy part - ha! :)
> 
> I say this only because I feel Agent Perl as it stands suffered from a
> lack of discussion & input.  Don't want to make the same mistake twice.
> 
> 
> 
>>If the Perl Agents web-site is no more, I'd be willing to host it here
>>at IT Carlow (assuming the current state of the site - HTML, etc.) can
>>be e-mailed to me).
> 
> 
> The website is still online at:
> 
> 	http://www.epn.nu/~spurkis/Agent/
> 
> Unfortunately, the README still says epn.ml.org, which closed down years
> ago.  As no new release was done, the docs haven't been updated :-/.
> 
> I've no problems with moving the site, though I'd recommend it wait until
> we know where this is headed.  Eventually I'd recommend registering
> agentperl.org, and setting up a CVS repository for development projects.
> 
> Glad to see people are still interested in Perl agents!
> 
> Regards,
> +--
>  Steve Purkis <spurkis@epn.nu>
> 
>