Re: MT draft

Ivan, thanks for the thorough and rapid feedback!

I will make some edits later tonight, but:

On May 10, 2013, at 5:48 AM, Ivan Herman wrote:

> Pat,
> 
> I have read through it and have a bunch of minor comments. I do not have any problem with the decisions you took (extensional semantics, etc).
> 
> I still wonder whether it would be helpful for the user to have all the rules, sorry, entailment patterns, collected as one table in an appendix (not instead, but as an addition to what you have); it makes an easier reference. Actually, the title of section 14 'Complete sets of entailment rules' is misleading; that section does not contain the complete set but, rather, describes the effects of using generalized triples...

Yes, "complete" is being used here in the logcial sense, which is probably a mistake. I will rephrase, and maybe also include the complete table as you suggest.

> 
> Here are the more detailed comments:
> 
> - As an editorial trick, I wonder whether it is worth using a javascript button that makes it possible for the user to hide/show the 2004 references, and/or the recondite entries (ie, the pink and blue texts). Same as in, say, the OWL 2 documents

I am leery of getting too fancy with things like this, but will check out the OWL2 sources.
> 
> - Both OWL and RIF have been re-published lately (OWL on 2012-12-11, and RIF 2013-02-05), both as '2nd editions'. The references should be updated.

I am constrained with references because of using ReSpec, so I think will need to ask Robin to update the biblio file. (Do you or anyone know, is it possible to insert references 'manually' when using ReSpec??)

> 
> - Section 3: "N-Triples syntax described in the Turtle Working Draft [TURTLE-TR]" -> "N-Triples syntax described in the N-Triples Working Group Note[@ref to http://www.w3.org/TR/2013/NOTE-n-triples-20130409/]"
> 
> However... is it correct that we use N-Triples? AFAIK, "25.0"^^xsd:real is not valid in N-Triples (it has to use full URI-s)...
> 

Indeed. I think I will simply say that I am using Turtle syntax. Nobody should be obliged to read all this stuff written using full IRIs.

BTW, I know the various examples are not done consistently, this will get fixed. 

> - Section 3, after the reference to [TURTLE-TR] (which should be now [N-TRIPLES]): isn't there a comma or a semicolumn missing? I have difficulties to properly parse the sentence.
> 
> - Section 3, on Issue 1: I think the question is whether a particular term is, or may be, useful outside the semantics document; for those, Concepts might indeed be a better place. I think this holds for 'name', 'vocabulary', 'empty graph', 'subgraph', 'ground graph'. I am less sure about the the various instances and the lean graph.

Yes, fair enough. 
> 
> - Section 6: I would put the paragraph 'This is clearly decidable...' into a blue box (ie, recondite content)

OK.
> 
> - Section 6: 'set share a blank node. then the set' -> 'set share a blank node then the set'
> 
> - Section 6: There is an extra line break in the paragraph 'Both graphs can be ...' before the expression for IEXT
> 
> - Section 6: I find the usage of '=='' for the various statements a bit strange in terms of typesetting. The 2004 version used a name for all of them and used bold characters of those names; that did look better (although finding right names may be a pain). Maybe a different font? A different bullet character? (Yes, I know, this is a matter of taste)

I agree == is a makeshift crock. I will find a snazzier way to do this in CSS but havnt got it done yet. 
> 
> - Section 9: Formally, L2V is not defined, though used in the Semantic condition.

Whoops. Will fix. 

> It is clear that L2V is a function mapping IR into the set of lexical-to-value mappings, but isn't it necessary to make that explicit for the definition? Although then the terminology 'If the literal is ill-typed then the L2V mapping has no value' is not, strictly speaking, true, it is the L2V(I(aaa)) that has no value... (yes, I am just picky)
> 
> - Section 9.1, the first paragraph refers to xsd:real, but the examples do not. I presume it should say "25.0"^^xsd:real in the first example, right?

Yes.
> 
> - Section 11: shouldn't the title be 'RDF-D entailment'?

Technically yes, but in fact the RDF-D is really only a precursor to the real point, which is to define RDF entailment. And I think that seeing RDF-D instead of RDF in the contents is liable to put some readers off, no?
> 
> - Section 12 (and subsections): shouldn't it be RDFS-D entailment everywhere? There is an insonsistency: the titles are without the '-D' but, for example, section 13 begins by 'S rdfs-D-entails E'...

See above.

> 
> - Section 12: My (admittingly non-native English) interpretation of the sentence 'a resource which represents a set of things in the universe which all have that class as the value of their rdf:type property' suggests that the given class is the only value for rdf:type. (I know that mathematically it does not say that.) Maybe saying 'as a value' instead of 'as the value'?

I will wrestle this back to clearer language. (Memo to self: no conditional clauses more than four deep...)

> 
> - Section 12: 'etc. . As' -> 'etc. As' (right after the 'some rdfs-valid triples' table)

Ah, I should check a style manual. I was always taught that the abbreviation dot was separate from the end-of-sentence dot. 

> 
> - Section 12.1 I think that the remark '(although it does contain the Unicode string ' "24"^^http://www.w3.org/2001/XMLSchema#integer')', though true, is just making it harder to understand. Better leaving it out imho.

Yes, I wondered about that. I thought it was cute, but it is hard to read. 

> 
> - Section 14, paragraph before the G-RDF-D table: 'generalized RDF triples, etc..' -> 'generalized RDF triples, etc.'
> 
> 
> Thanks!
> 
> Ivan
> 
> 
> 
> 
> 
> 
> On May 10, 2013, at 07:47 , Pat Hayes <phayes@ihmc.us> wrote:
> 
>> 
>> I just pushed a major edit of Semantics. It still needs work and cleanup before LC, but y'all might want to take a look to see if you approve.
>> 
>> Section 3, last para has discussion on denotes/refers versus identifies.
>> "Lemma" language has been eliminated.
>> Merge has been put back, and example given of union vs. merge.
>> Tables of "entailment patterns" have been inserted, but they aren't called "rules" in any normative text.
>> FInal informative section 14 mentions rule sets and generalized RDF syntax and states the terHorst completeness and complexity results, without proofs and treated descriptively rather than formally.
>> The 'extensional sematnics for RDFS' section has been eliminated. If we want it back, it should be another appendix. 
>> 
>> I havnt done the various appendices yet. 
>> 
>> I would like to shorten and simplify the current appendix C, or eliminate it entirely (?) Comments?
>> 
>> Pat
>> 
>> ------------------------------------------------------------
>> IHMC                                     (850)434 8903 or (650)494 3973   
>> 40 South Alcaniz St.           (850)202 4416   office
>> Pensacola                            (850)202 4440   fax
>> FL 32502                              (850)291 0667   mobile
>> phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
>> 
>> 
>> 
>> 
>> 
>> 
> 
> 
> ----
> Ivan Herman, W3C Semantic Web Activity Lead
> Home: http://www.w3.org/People/Ivan/
> mobile: +31-641044153
> FOAF: http://www.ivan-herman.net/foaf.rdf
> 
> 
> 
> 
> 
> 
> 

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes

Received on Friday, 10 May 2013 17:26:16 UTC