[SCL] SCL 1.02

[pat hayes]

>I've put a draft of SCL version 1.02 on the web:
>
>   http://cl.tamu.edu/docs/scl/scl-latest.html
>
>Sorry for the delay; the section on SCL and traditional first-order
>logic kept getting more and more interesting, and I ended up working out
>a lot of details.  It is fairly dense, but I think the ideas are clear.
>
>I'm a bit muddle-headed from lack of sleep at th moment, so there are
>bound to be some errors and other infelicities that I've missed.
>Comments/corrections/criticisms welcome.

Text bugs:
the 3rd paragraph in Lexicons refers still to PredVar.
implementated -> implemented (?)
The link from the footnote-5 mark is broken (needs a # in the href)
last para of SCL Semantics/Interpretations , too many "Qua"s

Quick suggestion: make arity a function from 2|(S*) -> 2|n whose 
value on a set of sequences is the set of lengths of those sequences. 
Then the 'conventional' constraint can be expressed by requiring 
arity(R) to be a singleton.

Another suggestion for presentation: give the lexical forms as a 
table, then define concrete syntaxes by displaying the concrete forms 
in a parallel table column. We can do this later, of course.

Trash the term 'relations-in-intension', and just say 'relations'. 
Only about ten people on the planet will know the first term and it 
will scare many readers away.

This syntax doesnt allow for terms in relation position.  Did we 
decide to disallow that case? There is a good case for disallowing 
terms in function position, but relation-valued terms seem OK to me. 
(?) I guess they complicate the SCL<->TFOL mappings.

I think we need to do some work on the conformance definitions - we 
will need for example to actually SAY what it means to claim that an 
inference engine is complete WRT an SCL language, because  some 
current specs define it differently than you or I would.  But lets do 
that later.   And there is a bit too much repetition of 'this is SO 
bloody obvious' in the last part :-).

We need to think about allowing other 'forms' in the syntax, eg 
comments, special character classes for typed domains, things like 
that. I don't have a feel for how extendable your syntax 
specification style is.

>-chris

PS Firebird has a really sexy code viewer :-)


-- 
---------------------------------------------------------------------
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 ihmc.us       http://www.ihmc.us/users/phayes