RE: Review of the RDF based semantics document (CLOSE ACTION-316)

Hi Ivan!

Many thanks for the review. Below, I will answer both of your messages in this single mail.

Please note that there is also a question to you (marked by "QUESTION").

Best regards,
Michael

**************************************************************
******************** First Review message ********************
**************************************************************

>-----Original Message-----
>From: Ivan Herman [mailto:ivan@w3.org]
>Sent: Tuesday, March 31, 2009 3:47 PM
>Subject: Review of the RDF based semantics document (CLOSE ACTION-316)
>
>Michael,
>
>Here are my review notes on the RDF Based Semantics document. All of
>them are editorial.
>
>Note that I should still read through the proof of the correspondence
>theorem, I presume, but I have to take a deep breath before doing
>that...:-)
>
>Ivan
>
>-------------
>
>
>General editorial/English question: as a mathematician I am used to the
>'iff' term. I wonder whether this is so universally known that it is not
>even necessary to define what 'iff' means...

Generally agreed. I even realized that the document used occasionally the long form "if and only if". Here is the way I have treated this:
 
* In the tables of Section 5 and their preceding explanation texts I have kept the "iff" in order to keep the middle column narrow, but I have added an additional item to the list of "conventions" at the beginning of the section, telling that "iff" in the section means "if and only if".

* In all other places (namely in the definitions and in the correspondence theorem), I have written "if and only if".

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21129&oldid=21127>

>------------
>
>Introduction, penultimate paragraph starting with "Significant effort":
>
>"[OWL RDF-Compatible Semantics]. the OWL " -> "[OWL RDF-Compatible
>Semantics]. The OWL"

Ok!

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21130&oldid=21129>

>----------
>
>Section 4.1, paragraph starting with "Note that it..."
>
>The text says:
>
>"The definition of datatypes with facets in Section 4.1 does not
>suggest..."
>
>but this _is_ section 4.1, so this is a self-reference...

Ok! (probably resulting from moving text passages around)
 
DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21131&oldid=21130>

>---------
>
>Section 4.1, paragraph starting with "Also note for..."
>
>"Also note for a datatype d and a facet-value pair < F , v > in FS(d)
>that the value" ->
>"Also note that for a datatype d and a facet-value pair <F , v > in
>FS(d) the value

Ok!
 
DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21132&oldid=21131>

>----------
>
>Section 4.1, paragraph staring with "In this document, it will always be
>assumed from now"
>
>The second paragraph seems to be superfluous, it repeats the same
>message...

QUESTION: Which "second paragraph" do you refer here? I do not find any redundancy in the text.

>-----------
>
>Section 4.2, second paragraph (definition of I),
>
>"provided that d is a datatype of D, I(u) = d, and" ->
>"provided that d is a datatype of D, IS(u) = d, and"
>
>Actually... I think a usual abuse of the syntax is to use the I(E)
>formulation for an interpretation when this means, mathematically, is
>IS(E) where 'IS' is the mapping defined in 'I'. As this shorthand is
>used all over the place, it might be worth noting it here.

I have deliberately chosen to use "I" instead of "IS", since 

* the RDF Semantics explicitly defines it this way in the "semantic conditions for ground graphs" (Section 1.4 of the RDF Semantics), and

* the RDF Semantics does use "I" instead of "IS" consistently in the whole document (see for example the table on "RDFS semantic conditions" in Section 4.1 of the RDF Semantics).

I don't want to deviate from the practice used in the RDF Semantics document without any good reason. One good reason would be that some nomenclature is used throughout the OWL 2 spec in a different form than in the RDF Semantics, but this is not the case here.

So I am not intending to change this.

>-----------
>
>Section 5, second paragraph
>
>"universe of the regarded OWL 2 Full interpretation" ->
>"universe for the OWL 2 Full interpretation being considered"
>
>(or something similar). The 'regarded OWL 2 Full interpretation' sounds
>funny to me...

Ok! There were other occurrences of this phrase, too.
 
DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21133&oldid=21132>

>there is also part of the sentence that seems to be out of place and
>probably unnecessary:

Actually, the sentence stops after "Section 4.4", so it was probably fine. But I had to look twice myself. :) So I have reworded the text slightly.

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21142&oldid=21133>

>---------
>
>Appendix 7, 3rd paragraph
>
>"and which must further meet all the global syntactical restrictions on
>OWL 2 DL ontologies that are specified in [OWL 2 Functional
>Specification]."
>
>First of all, the link seems to be wrong, it currently links to
>#ref-owl-2-rdf-mapping (and this is also true to the previous reference
>in the text). 

Ok! 

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21143&oldid=21142>

Note that I plan to check all the hyper links right before PR (I'll do this time consuming task only once).

>Furthermore, it is probably worth to explicitly link and
>refer to section 3 of the func spec, which collects all the dl specific
>restrictions in one place.

Ok, and so for other places in the document as well.

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21145&oldid=21143>

>Last paragraph in the same section
>
>"...while the new entailment query keeps being semantically equivalent
>to the original entailment query under the OWL 2 Direct Semantics."
>->
>"... and the new entailement query is semantically equivalent to the
>original entailement query under the OWL 2 Direct Semantics"

Ok!

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21146&oldid=21145>

>----------------
>
>Big example in the proof of the correspondence theorem:
>
>"corresponds to a union class expression in A. Since the pair &lang G1 ,
>G2 "
>
>I am not sure which XML entity you wanted to put for &lang...

It's "&lang;", which renders as "<". Fixed!

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21147&oldid=21146>

***************************************************************
******************** Second Review message ********************
***************************************************************

>-----Original Message-----
>From: Ivan Herman [mailto:ivan@w3.org]
>Sent: Wednesday, April 01, 2009 1:04 PM
>Subject: Addendum to my review of the RDF Based semantics (ACTION-316)
>
>Michael,
>
>I have also gone through Appendix 7, ie, the correspondence theorem. I
>cannot claim to have checked all RDF/Direct semantic condition pairs:-)
>but at least I understand what happens in the theorem and its proof
>based on the text and I did not find any issue...

Great, many thanks for the effort! :)
 
>My few (and purely editorial) comments are below
>
>Cheers
>
>Ivan
>
>
>--------------
>
>This is not strictly on Appendix 7, but is clearly related:
>
>Section 5, notes on the semantic conditions, first paragraph:
>
>"On the other hand, the RDF encodings of OWL 2 descriptions (Section
>3.2.4 of [OWL 2 RDF Mapping] and Sections 6 - 8 of [OWL 2
>Specification]), such as property restrictions,"
>
>In both the mapping and the syntax document the term 'expressions' is
>used and not 'descriptions'. For a better cross reference, the word
>'expressions' should be used here, too.

Thanks, this was only a typo. The document uses "expression" everywhere else.

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21148&oldid=21147>

>-------------
>
>Comments on Appendix 7 itself
>
>
>Stylistic issues:
>
>it may make the proof easier to read if some notations are defined
>upfront; that would avoid repeating them in the statements and the
>proofs. I am thinking of:
>
>  - G is an OWL DL ontology in Graph form and F(G) is its FS
>representation, meeting the restriction for DL and that results on the
>reverse mapping
>  - D is a datype for OWL FUll, and F(D) its Direct Semantic version.
>Actually, as an abuse of notation I do not think it would lead to any
>misunderstandings if F(D) was dropped in the sense that it should be
>made it clear in the introduction that the there is such thing as F(D)
>(as defined now) but then say that in the remaining of the section we
>use the same symbold with the understanding that Direct Semantics is
>meant in conjunction with F(D)
>
>These terms are repeated all over the place (which is, of course,
>mathematically correct) but makes the reading of the text fairly
>difficult.

Thanks! I will think further about this but I am not doing any work on this /now/.

At least the editorial quality of the proof is certainly not yet comparable with that of the rest of the document, because my main concern was to just have some *technically* satisfying proof before we go to LC. As you can see from the first EdNote at the beginning of the document, I am explicitly planning to further refine the proof, mainly in the time between the start of CR and until PR. I am keeping your comment at a save place till then. :)

>---------
>
>Point "C", in the balancing lemma proof, first bulleted item:
>
>"The removal of annotations by (a) and deprecation triples by (b) from
>G2*, as well as the replacement of the ontology header of G2* by (c)
>also do not hurt any syntactic restrictions"
>
>->
>
>"The removal of annotations by (a) and deprecation triples by (b) from
>G2*, as well as the replacement of the ontology header of G2* by (c) do
>not hurt any syntactic restrictions either"

Ok!

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&action=history>

>------------
>
>Proof of the Theorem, paragraph starting with "G1 and G2 are OWL 2 DL
>ontologies in RDF graph"
>
>"From this follows that the same literals..."
>
>->
>
>"This means that the same literals...

Ok!

DIFF: <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&diff=21151&oldid=21150>

--
Dipl.-Inform. Michael Schneider
Research Scientist, Dept. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: michael.schneider@fzi.de
WWW  : http://www.fzi.de/michael.schneider
=======================================================================
FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts, Az 14-0563.1, RP Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
=======================================================================