[SCL] Re: (SCL) Getting a move on

[pat hayes]

>Chris, Pat, et al.,
>
>There are several important requirements for CL/SCL:
>
>  1. Support existing users of languages such as KIF, CGs,
>     etc., with a semantic foundation that is upward
>     compatible (with some hopefully minor exceptions)
>     with what they have been using.
>
>  2. Present that foundation and its concrete embodiment
>     in usable logic notations in time to be used for
>     projects that have a pressing need for them ASAP
>     (where ASAP is approximately June 2003).
>
>  3. Do not create unnecessary impediments or constraints
>     that would undermine ongoing research into more general
>     foundations that could support powerful extensions
>     for the future.
>
>I believe that we can meet these requirements by producing
>a first version of a standard that can be presented to ISO,
>W3C, and others along the lines that Pat has proposed as SCL.
>
>I also believe that SCL could be defined in the more powerful
>CL foundation that Pat and Chris have been developing.  Therefore,
>SCL could be presented as v 1.0 next summer with a more general
>CL to follow as v 2.0 at some time after that.
>
>Re seqvars:  For full CL, we have been trying to avoid any
>ontological commitments whatever.  But for SCL, it would be
>possible to compromise the purity by including an ontology for
>integers and lists (which are needed in almost all applications).
>Then the lists could be used to support sequences as currently
>used in those KIF applications that need seqvars.  This approach
>isn't as elegant or powerful as the full CL proposal, but it
>would be a valuable stop-gap that would enable us to finish
>the definition of SCL by June.  Meanwhile, work on the CL
>standard for v 2.0 would demonstrate how SCL could be supported
>in a upward-compatible fashion.

John, I agree almost entirely and have been thinking along exactly 
the same lines. I think the built-in stuff needs to also include XML 
documents as a 'type', and maybe a couple of other things like XSD 
basic datatypes and uriref syntax. None of these are hard or 
controversial, but we need to keep them in mind when getting down to 
details.

I also think  that the ideas in the emails from Tanel are very much 
to the point and should be incorporated. Again, I dont think this is 
going to be hard to do.

In order to get this done in the timescale, by the way, we should 
probably try to get the main design of the language and the model 
theory done first, then think about concrete syntaxes, probably in 
subgroups focussed on each one. We will need at least a month to get 
the final documents written and debugged, and I think we may find 
that the detailed stuff takes longer than the abstract syntax. So we 
should try to get the first part done as quickly as possible. If we 
start with Chris' abstract syntax for CL and simplify it down from 
there, or at least work within that style, I think we will make the 
best progress.

Gotta go, time to raise seat backs to full upright position....

Pat


>John Sowa


-- 
---------------------------------------------------------------------
IHMC					(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola              			(850)202 4440   fax
FL 32501           				(850)291 0667    cell
phayes at ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam at ai.uwf.edu   for spam