W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > September 2001

Re: model theory publication draft

From: Pat Hayes <phayes@ai.uwf.edu>
Date: Thu, 20 Sep 2001 11:54:26 -0500
Message-Id: <p05101007b7cfc9bc7d33@[]>
To: w3c-rdfcore-wg@w3.org
Jeremy wrote:
>(Pat: my comments refer to I think not quite the latest draft, sorry if
>any of the issues have been fixed).

(I forgot to CC this reply to the rest of the group. The changes 
mentioned, plus some minor tidyings-up,  have been made in the 
version which has a CSS-style-removed line at the top  dated 9/20 
11:25 am, which y'all should now be able to download.)

>The Model Theory draft is stunning.
>I am certainly happy to see it go out.
>I particularly liked the touch of section numbering from 0.
>From the very first symbol in the table of contents it is clear that
>this aims to be a serious mathematical document, rather than a
>computer scientist's hack.

We mathematicians do have a certain sense of style. For example, I 
almost always put on a bow tie before typing long formulae.

>I have a few comments: three minor contentful points, and a few
>1: rdf:domain and rdf:range from RDFS spec.
>In Section 4, in the paragraph before the table, the phrase
>"rather than the wording in the M&S" should be
>"rather than the wording in the RDFS spec."


>2: Finiteness
>In section 5, in the paragraph after the table, there is the word
>This is the first mention of finiteness in the document, personally
>I was not assuming when reading this that the vocabulary was finite.
>I believe infinite RDF models are out-of-charter, so would be happy
>with making all vocabularies finite by changing the first paragraph of
>section 1.4 to start
>"All interpretations will be relative to a finite set of URIs, called
>the vocabulary of the interpretation"

Hmmm.  I removed the finiteness condition from an earlier draft 
because I wanted to allow infinite vocabularies as a theoretical 
construction, to allow for interpretations of things like *all* 
instances of certain URI schemas, or whatever.  But the only place 
where this was actually being used was in a section that was very 
speculative and has been removed, so we could put it back in.

On the other hand, the MT works fine for infinite vocabularies, so it 
seems a pity to restrict it unnecessarily; and even if such things 
are outside the charter, there are advantages to being constructively 
agnostic. How about if instead I simply qualify the reference in 
section 5, by saying that *if* the vocabulary is finite then the 
rules will terminate. (And if it isn't, then of course they won't, 
but we can still talk about their closure, its just that we couldn't 
compute it by running the rules to exhaustion.) (I've made that 
change in the latest version, if you do a reload in about an hour; 
its just after the table of rules.)

>More conservatively, finiteness could be introduced in the second
>paragraph of section 5
>"Suppose E is a finite RDFS graph "

See above.

>3: RDFS entailment rule missing?
>In the table in section 5, I think rule 3
>aaa rdfs:subPropertyOf bbb .
>    =>
>aaa rdf:type rdf:Property .
>should also have
>bbb rdf:type rdf:Property .
>on the RHS.
>This is only necessary when E does not contain any
>xxx aaa yyy .
>since otherwise the triple follows from the other rules.

Right, I missed that case. OK, done.

>Also, I would prefer if the rules were numbered.

Hah. They were originally, but I kept having to re-number them and I 
didn't use the numbers anywhere, so I trashed them.

I have put numbers back in the latest version.

>(Perhaps those who know me and have got fed up with me discussing
>grammar production 6.12, my favourite, would prefer if they
>4: Spellings ...
>Ones I noticed were:
>"sematnic" in 1.1
>"the the existential closure" in 2.1
>"decideable" in 3.1 after the Interpolation Lemma.

Thanks. I ran the text through three spellcheckers but they didn't 
find the the the. Fixed now.


IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
Received on Thursday, 20 September 2001 12:54:32 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:04 UTC