[SCL] some explanation

[Pat Hayes]

Mike, let me try to explain how the SCL ideas started to go a little 
beyond the CL=KIF+CGIF idea.

There are two related but slightly different notions of what 'common 
logic' might mean. One, the original, is the common part of a variety 
of surface forms of logic. That is, CL is what is common to a variety 
of notations, with the semantics attached nicely. But Ive always had 
a larger aim than (just) this. There is also another notion, of a 
logic which can be used in common by agents which are communicating 
logical content to one another (think of the semantic web at this 
point.) The first sense of 'common' is part of this, of course, since 
the agents might want to use different surface notations; but there 
is more to it than that. There is also the point that one agent might 
want to use a function where another wants to use a relation, or one 
agent's relation might be another agent's individual, etc.. What kind 
of logic could express what they have in common, and allow them to 
communicate? (i.e. in the same language, now: lets assume that we 
have solved the choice-of-surface-notation issue; this one still 
remains.) The wild west syntax and Chris' semantics solves this 
problem as well, very elegantly. And then there is yet another issue, 
which is what if one of them is assuming a different universe from 
another? Now their ranges of quantification don't line up. The SCL 
machinery, using named modules with headers, solves that problem too, 
using a small tweak of Chris' semantics. One relatively simple logic, 
designed carefully, solves all these problems at once. Moreover, 
talking about it in this way provides an elegant (seems to me anyway) 
justification for the design decisions: you can show that it kind of 
has to be the way it is, in order to fulfil its role as a 
freely-interoperable common logic - that is, a logic used in common 
across a communication network (not requiring reformatting or 
translation or syntactic negotiation, but can simply be sent from one 
place to the other, used freely by all agents using the same logical 
principles, morphed into a variety of surface forms, etc.; but still 
always the same, common, logic.)  Ive tried to write this up in the 
introductory stuff in the drafts you can read, and there are loner 
versions in some of the earlier drafts which I can point you at if 
you are interested. I believe that SCL is the first logic ever 
proposed that attempts to meet requirements like this.

So I've been on a kind of roll here for the past year, finding that 
the techniques that Chris and I (mostly Chris) invented to fix the 
first issue also solved these other problems (which I think are 
actually more pressing than specifying a common core for KIF and 
SGIF, in practice: the Sweb needs this kind of stuff): and the SCL 
drafts have been the product of this work. There are many other 
aspects that Ive been chasing down, eg how to incorporate public 
datatyping standards into a web logic, how to best embed RDF and OWL 
into it, etc..

Now, I think this would make a terrific proposal for a Web standard 
logic. But it does go beyond the original CL idea, and requires (to 
be done right) that the documents be organized differently. For a web 
logic, for example, the XML syntax is not just another concrete 
syntax, one option among many: it is the primary mode for information 
exchange, and so plays a central role. Hence my 
walked-into-a-brick-wall feeling today while talking to you and 
Harry, when I realized that you guys were not on the same wavelength 
as I was. I guess I had been assuming that everyone was kind of 
keeping up with the ideas in the SCL documents, but of course I 
shouldn't have. And as there isn't time now to reconvene the CL 
group, I will have to cut just back to the original idea for the ISO 
draft. Pity, IMO, but I guess as long as the new proposal containing 
these ideas (which is what this will have to be: I will write it up 
as a W3C submission) is consistent with an ISO standard then no 
serious harm will be done.

Pat
-- 
---------------------------------------------------------------------
IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
phayes at ihmc.us       http://www.ihmc.us/users/phayes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://philebus.tamu.edu/pipermail/scl/attachments/20050302/a21fc261/attachment.html