[SCL] Re: some explanation

[Pat Hayes]

>All,
>
>I was happy to see this part of the resolution of this thread, and I 
>hope everyone agrees with this.

I believe so, yes.

>
>John F. Sowa wrote:
>
>>On the following points, however, we have no
>>disagreement:
>>
>>  1. We will go with the new semantics that you [Pat]
>>     and Chris have developed, which is clearly
>>     a break with the old KIF semantics.
>>
>>  2. We will provide an XMLified notation in
>>     addition to the LISP-like and CG-like forms.
>>
>>  3. I also agree that there are new requirements
>>     that have been introduced to accommodate new
>>     developments in the WWW and elsewhere.  They
>>     will be met by the version you have developed.
>
>With respect to this last bullet, I understood Pat to say these 
>semantic  extensions are necessary but not sufficient for his 
>would-be "Web logic" aka "SCL The Next Generation".

I don't think that SCL-TNG is really necessary. The main parts of my 
'web logic' vision are now incorporated into the CL draft. The parts 
of SCL that are needed for this are the idea of giving a Web 
identifier to a piece of text , which as far  as the logic is 
concerned is just a name for a piece of text, and allowing what Ive 
called the 'module' construction , or something logically equivalent, 
to allow universe of discourse to be properly aligned, especially 
between different dialects. From the WWest syntax perspective, which 
is the appropriate one for any kind of Web-wide logic, modules are 
just syntactic sugar for restricted quantification and declarations 
of some names as being non-individuals (both of which are, in effect, 
already used in OWL, for example). The only wrinkle on Chris' 
semantics is a trick to allow the very same MT to apply both 
'globally' to modules seen from the outside, and 'locally' to modules 
seen form the inside, so that one doesnt have to say it all twice.

Now, a future SCL-TNG which provided for things like non-monotonic 
rules, modalities, meta-descriptions, etc..  might be a good thing to 
pursue, but it would be a new standard or a new extension  to the CL 
standard and would refer to it strictly, extending it rather than 
breaking it. Ive put a couple of unused syntactic plug-in points for 
just this purpose, to try to keep the likely future extensions 
aligned properly. Of course one can only think so far ahead, but Im 
optimistic that this is good for about a decade.

>
>I agree that the "Wild west syntax" is necessary and a good idea.
>
>I wonder what tweaks and twiddles with Chris's current semantic 
>model will become necessary for SCL-TNG.

Hopefully none, only extensions, see above.

>>I don't see any reason for not having identical
>>technical foundations for the ISO and W3C standards.
>
>I would agree with this, given that John and Mike accept the use of 
>SCL-as-we-know-it, with Chris's current semantics, as the common 
>model.  But in that case, I see no reason for there being two 
>standards -- at most we are talking about PDF and HTML versions of 
>the same document.

I don't think they could be the same document, since the W3C and the 
ISO treat matters such as conformance clauses (and the use of long, 
boring and completely uninformative glossaries) rather differently. 
It is a matter of house style. But the technical content can be 
aligned very tightly, with suitable cross-references.

>Further, assuming the "W3C standard" is really SCL-TNG,

No, the idea would for it to be the current version of SCL as far as 
possible, perhaps re-stated so as to give XML a more central role in 
the exposition (see below), but not with any logical changes.

>  the tweaks and twiddles may create a variance in the technical 
>foundation.  And retrofitting the SCL-TNG tweaks may have some 
>impact on the "conformance mappings" (the word we used to use was 
>"interpretation") of KIF and CGIF.

Right, this is exactly where the grit gets into the bearings. I think 
that with care and good communication (and some goodwill) on both 
communities, this can be avoided. The chief problem is simply 
communication. and I think that the informal but effective CL-mafia 
that has evolved now is capable of acting as a bush telegraph to keep 
several initiatives aligned.

>I don't yet understand Pat's concerns about the document organization

I had become enamored of an idea of using the XML syntax to play the 
role of an abstract syntax, since one can hang all the required 
metasyntactic structure on the XML itself. The result would be 
notationally ugly but I think extremely practical for implementors, 
and avoid what I have come to think is needless abstarction for its 
own sake.  However, this is now only of historical interest, as my 
XML view was overruled by Harry Delugach and opposed by others, and 
there was not time to argue the point. The ISO submission has 
therefore been written using an abstract-syntax style of 
presentation, and the XML syntax will be one of the concrete 
syntaxes. The practical advantages of the XML will be retained, 
however, so the issue was really only an editorial one about the 
structure of the document.

>, but I assume there is a related technical concern, if not for 
>SCL-as-we-know-it, then certainly for SCL-the-next-generation.  And 
>I suspect that SCL-TNG will require revision/replacement of a 
>particular section of the CL semantics, and therefore Pat would like 
>that to be a referenceable clause.  [Pat, can you make this issue 
>clearer?]

See above. However, the point you raise is an interesting one I had 
not considered, of how best to lay out the syntax and semantics so as 
to provide reference to parts of it, for future conformity 
statements. I think this deserves some careful thought.

>
>
>I strongly disagree with John, however, on the following:
>
>>But I believe that we should put the same technical
>>material into the W3C mill *first* so that they
>>have it on the web in a free version before it goes
>>through the ISO mill.
>>
>>ISO will then hold the copyright over the final
>>version that bears their stamp of approval, but
>>the technical material will already be available
>>anyway.
>
>If your community wanted a W3C standard, you should have gone there 
>in the first place.

We are going there, if I have anything to do with it :-)

>There is no W3C project for CL or SCL-TNG, and creating one will be 
>a political problem.  (And OBTW, Pat and Peter need all the help 
>they can get in the W3C political contest over logic languages.)

Indeed. However, we can but try. For the social affects that I am 
most interested in, W3C involvement is far more important than ISO 
approval. In fact, I have to say that it seems to me that the ISO 
should not even be in this business at all. But I accede to the 
enthusiasm that others have for ISO acceptance.

>Yes, you can individually submit your contributions to W3C as expert 
>contributions, and you can submit them to JTC1/SC32 as expert 
>contributions of the United States if you can get L8 to bless them 
>(which should not be difficult).  Either way they will be freely 
>available on the Web and have no status.

Status schmatus. What matters is whether people use them.

>I agree with Harry that we should continue the ISO project to the 
>point where we get a DIS and ISO limits distribution, and at that 
>point we apply to ISOCS for free dissemination of the standard 
>(which will be the following edition).  Along the way, it would be 
>advisable to find a target sponsor for the Web availability, and 
>that can be W3C or SemWebCentral or OMG or SourceForge, all of whom 
>represent different communities of probable interest.

Indeed.

>Make no mistake, we will be lucky to reach DIS by the Fall of 2005. 
>And logic standards are becoming a cottage industry, supported by 
>several "initiatives" with money (and the common realization that 
>XML itself is the surest route to the tower of Babel ever devised). 
>So we have to complete a consensus CL, before it becomes 
>yet-another-exchange-logic, instead of *the* reference standard.  So 
>let's not dilute the effort with "W3C versions" until we have some 
>version with status.

I think we should avoid the term 'version'. However, there is a clear 
distinction between CL, which Harry is chairing and has been 
submitted to the ISO process, and SCL, which arose from joint work 
between Chris, Tanel Tamet, myself and a few others. The SCL draft 
was the input to the first round of the CL-ISO process. Obviously it 
is to everyone's benefit to keep things as aligned as possible, but 
if the ISO process makes drastic changes to CL, that need have no 
effect on SCL.

Pat


>-Ed
>
>--
>Edward J. Barkmeyer                        Email: edbark at nist.gov
>National Institute of Standards & Technology
>Manufacturing Systems Integration Division
>100 Bureau Drive, Stop 8264                Tel: +1 301-975-3528
>Gaithersburg, MD 20899-8264                FAX: +1 301-975-4694
>
>"The opinions expressed above do not reflect consensus of NIST,
>  and have not been reviewed by any Government authority."


-- 
---------------------------------------------------------------------
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