RE: Review of Direct Semantics (ACTION 314)

Hello Michael,

Thanks a lot for your review. Please find my answers inline.

Regards,

 Boris

> -----Original Message-----
> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org] On
> Behalf Of Michael Schneider
> Sent: 05 April 2009 23:45
> To: W3C OWL Working Group
> Subject: Review of Direct Semantics (ACTION 314)
> 
> Hi!
> 
> This is my review of the Direct Semantics. This document has already been in a
> good shape months ago, when I first reviewed it, and so it is now. I have only
> found smaller things, none of them critical.
> 
> Note: I did not check against Thomas' review that arrived a bit earlier, so
> there may be overlap.
> 
> * General: The definitions in the document, in particular those in §2.5, are
> of the form "A if B". While this is a typical convention under mathematicians,
> our documents are targeted to a broader audience. In order to avoid confusion,
> I suggest to always say "if and only if" (or "iff", and say once that this
> means "if and only if").
> 

I've had extensive discussions about this with many people (notably Uli), and they insisted that the latter form is rather ugly. I'd prefer leaving things as they are. 

> * Abstract: Suggestion: Consider not talking about inference problems in the
> abstract, but rather move that sentence into the Introduction. It is a pretty
> technical topic, and I think the abstract should only give a concise and clear
> hint about the /basic/ purpose of the document.
> 

I'm not sure about this: the purpose of an abstract is to summarize what the document does. In this case, it actually does define the inference problems, so I do not see why this shouldn't be mentioned in the abstract.

> * §1, 1st par: There is a reference to the Document Overview. I think this is
> redundant, since the Abstract already mentions it (in the part common to all
> documents).
> 

Agreed -- I've thrown it out.

> * §1, 2nd par: "Since OWL 2 is an extension of OWL DL, ..." Didn't we want to
> turn away from this statement?
> 

But this actually is true: each OWL DL ontology is an OWL 2 DL ontology, and it is therefore an OWL 2 ontology. Hence, I don't really see a problem with this statement.

> * §1, 3rd par: "The semantics is defined for an OWL 2 axioms and ontologies,
> [...]". Something is grammatically wrong with this sentence.
> 

This was already addressed by Thomas's review.

> * §1, 4th par: "OWL 2 allows for annotations of ontologies, [...], and other
> annotations." The last part of the sentence sounds confusing (to me). I guess
> you mean "annotations of annotations". If yes, then why not simply say so?
> 

I've rephrased this sentence as part of my answer to Thomas's review. Please let me know should you not be happy with this.

> * §2.1, before the list concerning datatype maps: The sentence ends with "with
> the following components.", i.e. with a ".". In other places, e.g. the
> following list for "vocabularies", there is a ":" instead. Consider being
> coherent.
> 

I've changed "." into ":".

> * §2.1, list of items (and other places): I always wondered why the pairs,
> such as "< F v >" and "< LV DT >" do not contain a comma, while other pairs
> later in the document do have a comma? Consider being coherent.
> 

Fair enough -- I've added a comma everywhere. I've also updated the notation in the Syntax document.

> * §2.1, datatype maps, last item: I find the name "facet value" somewhat
> confusing, since (a) it is actually a set and (b) I rather would expect the
> "v" in a pair "< F v >" to be called the "facet value". I wonder if there is a
> better name for this, but have to admit that I do not really have an
> alternative. If this is going to change, than other documents might need to be
> changed, either.
> 

Fair enough. I've actually avoided giving this thing a name and have just said that the interpretation function assigns to < F v > a set (< F v >)^FS. I've also changed the Syntax document accordingly. (There, only the heading of Table 4 needed to change.)

> * §2.1, vocabularies, 3rd item: Typo: "owl:bttomDataProperty" (missing "o").
> 

Thanks!

> * §2.1, vocabularies, 4th item: Are you sure that V_I contains anonymous
> individuals, which are existential variables?
> 

Yes. This is not a problem, however: we require the anonymous individuals in Ax to be standardized apart and we have appropriate restrictions in the definition of models. Introducing yet another syntactic construct would just complicate matters.

> * §2.2.2: "An n-ary data range DR is interpreted as an n-ary relation (DR)^DT
> over Δ_D." Please be more explicit what this means, just as you are more
> explicit earlier in this paragraph on what "unary relation over DELTA_D"
> means, namely "(DT)^DT subset DELTA_D". You probably mean "(DR)^DT subset
> (DELTA_D)^n"?
> 

Thomas had a comment about this, and I've rephrased the paragraph in response to him. Please let me know should you find the new formulation insufficient.

> * §2.5, "decidability": "Each class expression of type [...] can contain only
> object property expressions that are simple [...]". I wonder why there is such
> a detailed statement here, and why there isn't just a simple pointer to the
> Structural Spec? Isn't this duplication of information?
> 

I'm afraid this is really needed. Note that we define simple properties w.r.t. a single ontology. Here, however, a property needs to be simple across *two* ontologies. That is, you could have a situation where an object property OP is simple in O and O' alone, but this is not the case if you look at O and O' together. For ontology entailment to be decidable, however, you need the latter condition to hold; for example, you should not use in O' a property in a number restriction if that property is transitive in O. Thus, this is something that cannot be checked by looking at a single ontology (which is what we do in the Syntax document); rather, it is a "local" property of ontology entailment.

> * §3: I'm unclear whether this section is intended to be normative or not? I
> would rather say that it is informative.
> 

Indeed. I've changed the heading to indicate this.

> * References: The [OWL 2 Profiles] reference doesn't terminate with a ".".
> 

Fixed.

> Cheers,
> Michael
> 
> --
> 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
> =======================================================================
>

Received on Monday, 6 April 2009 13:14:18 UTC