Sandro,
I'd like to see the CL abstract syntax in asn06;
http://www.w3.org/2005/rules/wg/wiki/asn06
IKL uses one of the CL concrete syntaxes.
These are lisp s-expression syntaxes,
but it's annoyingly different from ACL2 syntax;
e.g. in ACL2, (and) gets case-folded to (AND),
but IKL makes use of the distinction.
http://dig.csail.mit.edu/2006/Papers/owl-acl2/doc20
The attach suggests there's some Java stuff that
might be isomorphic to CL-in-asn06. Here's hoping
you find time to take a look...
p.s. I picked public-cwm-talk somewhat arbitrarily;
RIF overwhelms my inbox, but I could use www-rdf-logic
or www-rdf-rules, I suppose.
--
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541 0875 0F91 96DE 6E52 C29E
Forwarded message 1
Hi John
On May 1, 2007, at 11:49 , John F. Sowa wrote:
> Bill,
>
> That's a good idea.
>
>> Comments welcome
>
> First comment: What's the URL of the web site?
I just submitted the application to SourceForge, so no URL yet.
> Second: The key to a successful joint project is some "seed"
> or "core" that other people can download, play with, and add to.
> Do you or your group have or know of any such core that could
> be used? Are there any KIF tools that would be suitable as
> part of a CL core? Which ones?
Glad you asked. I have on my machine a Java object model of CL
abstract syntax – similar to a DOM for XML. I will upload that as
soon as the project is approved. I will follow with a set of classes
for CLIF and then probably a converter from the CLIF model to
abstract syntax - of course conversion in the other direction is not
unique, but could be done with "style preferences" or something. If
I do a reasonable job of that (I'm not God's gift to Java
programming, but I'm ok) then it can serve as a template for other
concrete syntaxes, like CGIF and CLCE.
Once we get the base object models nailed down all kinds of things
could be done, like in-memory and disk-based stores, translators from
OWL and RDFS to CL abstract syntax, etc. I guess I'm hoping the end
result will be something similar to HP's Jena for CL – in fact it
would be a good idea to use Jena or something like it (if exists) to
handle the W3C stack portions.
> Third: At VivoMind, we have been developing CLCE (Common
> Logic Controlled English) as a front-end to CL. Our current
> implementation is proprietary because it's tied to other
> VivoMind projects.
You should be able to factor out the class representation of CLCE,
no? But I realize that's easier said than done sometimes. OW is
moving to ISO CL compliance (in particular with an extension of a
CLIF sub-language), so we're pretty much starting fresh -- for that
reason I figure there's no harm in putting up the code
> But we would be willing to make the
> grammar nonproprietary.
Great. Better yet if you have a grammar for a parser-generator like
ANTLR. We'll probably put up an ANTLR grammar as soon as we nail
down the CLIF object model to act as a target.
> Perhaps some funding could be found
> to implement at least the core of an independent, nonproprietary
> implementation.
_______________________________________________
CL mailing list
CL@philebus.tamu.edu
http://philebus.tamu.edu/mailman/listinfo/cl