[SCL] Telecon and SCL progress

[Tanel Tammet]

Hi,

pat hayes wrote:

> Greetings, and sorry for the delay in getting started.
>
> This email is an attempt to summarize the progress we think we made at 
> the Santa Fe ISO meeting a few weeks ago on the basic language, and to 
> outline the various things that need to be done. Hopefully we can 
> schedule a telecon for next week (this week I will be at the W3C 
> plenary meeting) to start discussing these.
>
> TELECONS
>
> Let us try to start having a telecon. We can schedule the second one 
> after we see how the first one goes. Early in the day US time seems 
> best for European members, so I suggest that we try for 0900 CT/1000 
> EST/1600GMT/1700 CET, on Monday 10 March.  If anyone finds this 
> time/day impossibly inconvenient, please suggest another.  The number 
> to call is (US) 413-351-4140. Please say your name when you arrive so 
> we can keep track of who is on the call. I expect the original 
> volunteers to phone in if possible, but others are welcome if they 
> feel they can contribute. The basic rule is that anyone I can talk to 
> is liable to be asked to actually do something.

This time is fine with me. Somewhat later would be a bit inconvenient, 
but possible. An hour or half an hour earlier would be optimal, but this 
might
be too early for US. Five-six hours later would also be OK.

> .... volunteers to take charge of drafting each particular concrete 
> syntax proposal. As a preliminary suggestion:
>
> XML: Tanel Tammet
> CG's: John Sowa
> KIF-style: Pat Hayes
> Principia-style: Chris Menzel ?
> RDF triples: Pat Hayes
> Z: John Sowa ?

OK for me.

.....

> 5. ANNOTATIONS AND COMMENTS SYNTAX
>
> Several people have suggested the idea that the syntax should provide 
> for text to be associated with subexpressions which has no logical 
> content. We need to decide on a way to do this. I would suggest that 
> we do this by allowing in the abstract syntax a 'wrapper' 
> construction, where a wrapper is a form with two parts, a text string 
> and body which is an expression of any syntactic form, and the 
> syntactic class of any wrapper is the same as that of its body. We can 
> then provide a distinctive concrete syntax for wrappers (eg square 
> brackets in a KIF-style syntax, or a particular XML tag in an XML 
> syntax) which would allow us to insert text at any point in any 
> expression. This allows wrappers to be nested.
>
> Issue: is it sufficient to only add text? Eg in RDF, comments may be 
> arbitrary expressions and may share variables with other expressions.

Plain text is universal, unlimited, and safe.

Anybody can use text in any specific ways by defining their own 
additional conventions for text.
I see no need to attempt to guess these conventions and then put the 
selected/guessed ad-hoc conventions into base language.
Since we treat FOL, IMHO we should try to avoid defining additional 
languages in the process.

> ....

> Issue: this requires a systematic way to refer to ontologies. Do we 
> need a syntax for naming ontologies? 

While I am pretty sure that naming ontologies, rules, etc are all really 
critical issues in practice (ie SCL would be
of very limited use unless metainformation of various kinds can be 
attached), I also think that in case we can attach text to
expressions, it is later possible to define conventions for the text 
which make all this possible. These conventions would be
at a separate level. Once we are sure that we provide good hooks for 
additional information, IMHO we should not be
tempted to immediately design something to be hooked up at once. If we 
do that anyway (might be good
for demonstration purposes, as an example of how things can be done), 
then we should very clearly
keep it at a separate level, unrelated to the basic language.

Regards,
           Tanel Tammet