[cl-semantic-devel] Choice of Lisp

[Alan Ruttenberg]

Another thing to consider would be to associate this effort with the 
Mindswap Pellet and SWOOP projects. They have a high quality OWL 
reasoner written in Java, and SWOOP, an ontology editor with some 
innovative features. We could incorporate abcl as an extension language 
(abcl is written in java and can easily call out to java and vice 
versa), and then build on their pretty strong foundation, which already 
has a GUI (though it could be made much better, I think). I've 
interacted with them somewhat and they are a really bright group of 
people.

Pellet: http://www.mindswap.org/2003/pellet/
SWOOP: http://www.mindswap.org/2004/SWOOP/
abcl: http://armedbear.org/abcl.html

I have prototype common lisp code using abcl which talks to Pellet that 
I'd be happy to share. There are a (very) few words about it at 
http://biopaxwiki.org/cgi-bin/moin.cgi/Using_Pellet

BTW, to introduce myself: I'm Alan Ruttenberg. I work in the 
computational biology department at Millennium Pharmaceuticals, for the 
most part focusing on pathway databases and methods to apply structured 
knowledge to the analysis of experiments, mostly transcriptional 
profiling. In the last year I've been more involved in semantic web 
related activities. I've used Wilbur, but more recently am using abcl 
with Pellet because it is a (free and open) OWL reasoner that works 
well.  My main interest in in BioPAX, a data exchange for biological 
pathways, though I'm also active in the W3C Semantic Web Health Care 
and Life Sciences Interest Group, and in correspondence with the OWL 
developer community.  I participated in last year's OWL Experiences and 
Directions workshop, with this paper:  Experience Using OWL DL for the 
Exchange of Biological Pathway Information 
(http://www.mindswap.org/OWLWorkshop/sub37.pdf) and will be on the 
program committee for next year's workshop. Don't know if my interests 
and code will mesh with what you are doing, but I'm interested in 
seeing where you go with this.

I'd be interested in hearing introductions from other folks who are on 
the list...

Regards,
Alan

On Feb 8, 2006, at 5:27 AM, Brandon Werner wrote:

> I was thinking we would want to maintain at least some ability to do a 
> RacerPorter type application, or to display data in a dynamic way. 
> Considering the current newness of AJAX, I'm not sure if that would be 
> an option, although I do believe the Java hooks would be preferable to 
> a Tk interface. 
>
> Also, I have added some menu items to the website, as well as written 
> half the installation page (targeting Allegro). I do agree that we 
> should be agnostic (I am writing setup for SBCL and Allegro), but I 
> thought it may be best to have one solid distro we would all agree 
> was crucial for the application to run in. I really don't want it to 
> be a heap-limited commercial distro.
>
> However, chances are do to speed and features it will be ACL.
>
> Brandon Werner
> brandon.werner at mac.com
> http://www.brandonwerner.com
> P +01 859 640.7148
> Skype: bbjwerner
> iChat - AIM: bbjwerner
>
>
>
>
> On Feb 8, 2006, at 2:01 AM, Alan Ruttenberg wrote:
>
>> Why would you choose a particular lisp as opposed to aiming for 
>> platform neutrality? (I use openmcl, abcl, sbcl, cmucl) There are a 
>> variety of compatibility packages.
>>
>> It would seem to me that this is a more sensible course. Also you 
>> might want to consider McClim, lambda GTK, calling out to java 
>> libraries, or AJAX technology for the UI all of which are a little 
>> more current than ltk. Unless you are thinking about some extremely 
>> cpu intensive user interface, I'd think it would make sense to go 
>> with a browser based solution.
>>
>> Regards,
>> Alan
>>  
>>
>> On Feb 7, 2006, at 11:11 AM, brandon.werner at mac.com wrote:
>>
>>> All:
>>>
>>> There has been some debate among the researchers about which CL to 
>>> use for this project. Although I have experience with both SBCL and 
>>> Allegro, it would appear that Allegro is not going to release their 
>>> 8.0 evaluation version anytime soon and this is the only version 
>>> that includes the asdf: functionality we will be using to maintain 
>>> and install builds of cl-semantic.
>>>
>>> Yes asdf can be easily added to Allegro, but SBCL is very easy to 
>>> setup, and in a project where are are attempting to get as many 
>>> developers interested as possible, many whom may not have university 
>>> access to Allegro, it seems important that we have a CL that is 
>>> ready right out of the box (we may even group all the source for 
>>> Wilber and other necessary requires in to one .gz for easy 
>>> deployment in to each developer's repository).
>>>
>>> However, from my experience using SLIME with Allegro and SBCL, SBCL 
>>> is extremely slow and is generally buggy on OSX do to Apple's Crash 
>>> Reporter actively looking for Mach exceptions because Mac OSX has no 
>>> idea that a program can do a SIGSEGV and keep on living.
>>>
>>> Anyone have any suggestions?
>>>
>>> I would also like suggestions on a GUI environment for cl-semantic, 
>>> perhaps ltk (uses Tk).
>>>
>>> Also, please start posting directly to the message group instead of 
>>> to individual email addresses so that we can leverage our new home 
>>> at common-lisp.net. We have had many people joint the development 
>>> and announce list and we owe it to them to let everyone know what is 
>>> going on, even if they have yet to join the project.
>>>
>>> Thanks!
>>>
>>> Brandon Werner
>>> cl-semantic project administrator
>>> http://common-lisp.net/project/cl-semantic/
>>> brandon.werner at mac.com
>>>
>>> cl-semanic is a collection of RDF/OWL extraction and relationship 
>>> parsing macros written in Common Lisp.
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> cl-semantic-devel mailing list
>>> cl-semantic-devel at common-lisp.net
>>> http://common-lisp.net/cgi-bin/mailman/listinfo/cl-semantic-devel