[racket] mobile agent in racket

From: Neil Van Dyke (neil at neilvandyke.org)
Date: Wed Jan 5 03:17:15 EST 2011

If you wanted to build a new mobile agents framework for research 
purposes, I think that using Racket would be viable.

Some relevant features that Racket gives you atop what typical Scheme 
dialects/derivatives do include: greater ability to limit resources, the 
serializable continuations work for the Racket Web Server, good support 
for defining and mixing languages (or constructing eval environments, if 
you like).

Also, if you are doing your academic research on a small niche like 
mobile agents, you will especially want to get connected with the right 
research circle.  Over ten years ago (when I last kept up with the 
topic), the mobile agents community was small, and there was a private 
email list shared among key researchers, but the email list seems to 
have moved ("mobility" pun) or disappeared.  I don't know whether work 
on this topic is still active, or whether the topic has morphed into 
something else.  I suggest finding and reading all the conference and 
workshop proceedings, as well as trying out any prototypes and tools you 
can find.  After you've ingested all you can find, you might try 
contacting a couple of the people whose work most interests you, to 
express your interest and ask for advice on how to proceed.

In addition to the topic of "mobile agents", also spend some time 
getting background on "distributed objects", if you have not done so 

Oh, before you do too much more reading and playing with software, 
perhaps you should write down your perspective on mobile agents and why 
it interests you, as notes to yourself.  Then you can refer back to 
those notes after being overwhelmed with other people's perspectives, so 
that you don't forget ideas that others might have missed or failed to 
follow through on.  (It is not uncommon for lots of people to have 
similar grand visions of a research topic, and for there to be lots of 
hype on a topic for a while, but then the hype dies down, and those 
people graduate and/or switch to new topics before they've achieved the 
grand vision for the previous topic.  That doesn't mean the previous 
topic is uninteresting or insoluble, nor that the dominant approaches 
during the hype were the best ones.)

Also, once you've scouted out the topic, be sure to get an academic 
research advisor, if you don't already have one.  Besides topical and 
general research direction, s/he can advise you about funding.

Good luck.

Nevo wrote at 01/05/2011 02:04 AM:
> hi,
>   I'm new to mobile agent, but i'm interested in that and plan to 
> continue my study on that if possible. Now I'm required to prepare a 
> proposal related with mobile agent and I want to use racket as the 
> implementation language, but i'm not sure if this is really 
> applicable, so i'm asking here for some suggest.
>   As the background reading, I've been briefly read papers "Scheme 
> implementation of the functional language for mobile agents with 
> dynamic extension 
> <http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1555148>" and 
> "Language Support for Mobile Agents, 
> <http://www.cs.virginia.edu/%7Eknabe/dissertation.ps.gz>". Both 
> implementation are based on functional programming language, so I feel 
> it should be feasible in Racket as well. After further searching, it 
> seems IBM's "Aglets <http://www.trl.ibm.com/aglets/>" is the most 
> popular implementation, in Java, and Python also has a prototyping 
> framework MAF <http://maf.sourceforge.net/> which supports "foreign 
> agent", and  Pyrover <https://www.taupro.com/ss/pyrover.html> which I 
> cannot open the link for (just read the cached page). I'm wondering is 
> there any similar framework or implementation in scheme or racket I 
> can reference from ? Thanks!
> Nevo

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20110105/bc0a6a23/attachment.html>

Posted on the users mailing list.