- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Thu, 09 Apr 2009 13:02:17 -0400 (EDT)
- To: schneid@fzi.de
- Cc: public-owl-wg@w3.org
I think that this is generally a good idea. The changes are almost entirely in the RDF-Based Semantics document. A few concerns: 1/ Trivial - There are two cases where OWL 2 Full vocabulary is written as "vocabularies for RDF, RDFS or OWL 2 Full", which will need to be checked carefully. 2/ Minor - The Introduction will probably need some wordsmithing after the term replacements have been done, particularly where it refers to OWL 2 Full ontologies. 3/ Moderate - "OWL 2 vocabulary" is not a phrase that I would use here, without some special preparation, even though "OWL vocabulary" was used in S&AS. In particular I would not go along with The OWL 2 vocabulary is a set of IRIs [RFC 3987], which occur in the sets of RDF triples that build the RDF encodings of all the OWL 2 language constructs [OWL 2 RDF Mapping]. I suggest instead using "OWL 2 RDF-graph vocabulary", I guess, although that is rather pedantic, particularly as it would be a frequently used term. To reduce the level of pedantery the document could define "OWL 2 RDF-graph vocabulary" and then say "(commonly abbreviated in this document to ''OWL 2 vocabulary'')". If the sentence above was then removed from the Introduction and the redundant sentence The OWL 2 Full vocabulary is a set of IRIs [RFC 3987] with the standard prefix name owl (see Section 3.1). was removed from 3.2, things would be cleaner. peter From: "Michael Schneider" <schneid@fzi.de> Subject: Occurrences of "OWL 2 Full" in our documents Date: Thu, 9 Apr 2009 14:46:42 +0200 > Hi all! > > Here is a sort of "Last Call". :) It's very late in the day, I know, but > this is a last attempt to cope with the current "OWL (2) Full" naming > situation. I'm still unhappy with it, and I am probably not the only one > (see below). > > > Summary: > -------- > > I will propose a set of renamings that will reduce the number of > occurrences of the term "OWL 2 Full" to a minimum in our documents, > while at the same time will clarify things considerably and, I believe, > will have the chance to satisfy all parties (inside and outside the WG) > under the existing constraints. I also believe that it will give us, the > WG, a stronger position in the next LC round (and afterwards), if > further criticism arises. The implementation of the proposal can be > performed rapidly, so there will be no risk for our publication > schedule. > > > Some Background: > ---------------- > > I recently saw Ian moan: > > <http://lists.w3.org/Archives/Public/public-owl-wg/2009Apr/0089.html> > [[ > I have updated the document to mention both > OWL 2 datatype maps (as defined in SS&FS) > and OWL 2 Full datatype maps (as defined in R-BS) > [...] > It is slightly unfortunate that this means more use > of "Full" -- I wonder if OWL 2 Full datatype maps > shouldn't have been called OWL 2 RDF-Based datatype > maps, but that is another issue. > ]] > > And Ivan had a point when reviewing the RDF-Based Semantics, and > apparently wasn't really perfectly happy with my answer: > > <http://lists.w3.org/Archives/Public/public-owl-wg/2009Mar/0555.html> > [[ > Michael Schneider wrote: >> >> Ivan Herman wrote: >>> >>> - there are some places where you refer to >>> 'OWL 2 Full Interpretation/Satisfaction/Entailement'). >>> Strictly speaking, this may not be 100% o.k., >> >> I called them this way, because you will find >> the analog names ("OWL Full XXX") in >> Section 5.3 ("OWL Full") of the SAS: >> >> <http://www.w3.org/TR/owl-semantics/rdfs.html#5.3> >> > True. But one of the discussion items we had in the > past few days was based on the recognition that there > was some sloppiness there... > ]] > > And I alway was (and am) disgusted by the current terminological chaos > in the RDF-Based semantics document. There is a mixup of uses of > "RDF-Based" and "OWL 2 Full", where "OWL 2 Full" is often used in a > context when only the semantics is meant. There are actually more than > 100 uses of "OWL 2 Full", but it never becomes clear what this "OWL 2 > Full" is, and why it is used in the respective place. > > Now, I'm really eager to make all of us happy(er)! :) > > So let's see what I believe we can settle on, or what we already have > settled on in the past: > > > Uses of "OWL 2 Full" that seem to be accepted > --------------------------------------------- > > (1) There are "OWL 2 Full ontology documents", as defined in the > Conformance document [1]. This term has been used in that document for a > long while now, and I don't remember that it has ever been put strongly > in question. > > (2) After some longish discussions, we seem to have settled on "OWL 2 > Full" meaning the combination "RDF-Based Semantics + RDF Syntax", as it > was (most probably) be meant in OWL 1 (although never precisely stated > there). We have claimed this in a response to an LC comment [2], and it > is actually said so in the Document Overview [3]. > > (3) The RDF-Based Semantics document introduces the term "OWL 2 Full > ontology" [4] as a synonym for "RDF graph". It seems to me that this is > fine (and straight forward), since we also use the term "OWL 2 Full > ontology document" (see item (1)), which refers to a concrete RDF > serialization, though. > > If we agree that these uses of "OWL 2 Full [xxx]" are fine, then I am > making the following proposals: > > > Concrete Proposal > ----------------- > > (A) Terms like "OWL 2 Full interpretation" and "OWL 2 Full datatype > map", etc. (see the complete list below), are renamed into "OWL 2 > RDF-Based interpretation" and "OWL 2 RDF-Based datatype map", > respectively, since they all relate to the semantics only. > > (B) Specifically, the term "OWL 2 Full vocabulary" [5] is renamed into > "OWL 2 vocabulary". This is fine, since the RDF vocabulary terms are > used in the RDF mapping as well, so they is not specific to the OWL 2 > RDF-Based Semantics. I have only named it "OWL 2 /Full/ vocabulary" for > naming consistency reasons. With proposal (A), these reasons will be > removed. > > (C) The name "OWL 2 Full" will be used in exactly one normative context > in the RDF-Based Semantics, but is introduced as a terminological > convention: In Section 2.1, which is about Syntax (RDF Graphs, aka "OWL > 2 Full ontologies"), in the paragraph that sais what an "OWL 2 Full > ontology" is ("A word on nomenclature") [6]. There, it will be said what > "OWL 2 Full" is (according to Item (2) above). > > (D) The (informative) "Introduction" section will also use the word "OWL > 2 Full" exactly once, when it talks about Section 2.1 [7]. So, it's > effectively the same use as in (C), but in a somewhat more prominent > position. > > (E) The "Changes from OWL 1 RDF-Compatible Semantics" section [8] > concerning the new naming conventions will say a few words about the > changes. This will primarily explain why a formerly "OWL Full > interpretation" is now an "OWL 2 RDF-Based Interpretation": Because it > makes sense and clarifies things. It will also be mentioned that the old > spec was pretty vague in what "OWL Full" meant, and that this has now be > clarified. > > > Discussion > ---------- > > This set of proposals will reduce the number of "OWL 2 Full" occurrences > from more than 100 to exactly 4 in the RDF-Based Semantics, and all of > them essentially in the same context. Further, in the Conformance > document, there will at least be no "OWL 2 Full datatype map" anymore. > > No one can accuse us then that we try to somehow circumvent the use of > "OWL 2 Full", or make it fuzzy, or whatever. It's just that the > RDF-Based Semantics document most of the time talks about, well, > semantics, hence the rare use of the name. But there *IS* one place > where it explicitly sais what "OWL 2 Full" is, and this will (must) > satisfy people. > > The only problem that might occur in LC is that there are people > insisting on the reuse of the legacy terms such as "OWL 2 Full > entailment". But we can answer that there was a tradeoff between clarity > and legacy (enforced by LC1 !), and we decided for clarity, since our > job is to write a spec. I would even offer to take such an LC response > up to myself! :) > > > Complete list of renamings in the RDF-Based Semantics: > ------------------------------------------------------ > > * "OWL 2 Full ontology": NO CHANGE! > * "OWL 2 Full vocabulary" --> "OWL 2 vocabulary" > * "OWL 2 Full datatype map" --> "OWL 2 RDF-Based datatype map" > * "OWL 2 Full interpretation" --> "OWL 2 RDF-Based interpretation" > * "OWL 2 Full (satisfaction|satisfies)" --> "OWL 2 RDF-Based > (satisfaction|satisfies)" > * "OWL 2 Full (consistency|consistent)" --> "OWL 2 RDF-Based > (concistency|consistent)" > * "OWL 2 Full (entailment|entails)" --> "OWL 2 RDF-Based > (entailment|entails)" > * "OWL 2 Full semantic conditions" --> "OWL 2 RDF-Based semantic > conditions" > > > (Potentially) Other Affected Documents: > --------------------------------------- > > * Document Overview: NO CHANGE, AFAICS! :) The only occurrence of "OWL 2 > Full" in Section 2.3 will just fit my proposal. > > * Conformance: > ** No change concerning "OWL 2 Full ontology document". > ** Rename "OWL 2 Full datatype map" to "OWL 2 RDF-Based Datatype > Map". Then, the OWL 2 XXXXX Datatype Map corresponds to the OWL 2 XXXXX > Semantics, which is fine. > ** Somewhat unclear about "OWL 2 Full entailment checker", since this > term is never used in the RDF-Based Semantics. My first idea was to > rename it, since it is about entailment checking, i.e. is semantics > related. However, then the other checkers would consequently also need > to be renamed, and this would lead to three "different" "OWL 2 Direct > entailment checkers" (for DL, EL, QL), and another "OWL 2 RDF-Based > entailment checker" (RL). Also, there is also a direct correspondence > between an OWL 2 XX entailment checker and an OWL 2 XX ontology > document. This looks quite useful to me. So I suggest to keep the name > as it is. > > * NF&R: Looks to me that nothing has to be done there. There is only a > reference to "OWL 1 Full", and this looks harmless to me [FIXME]. > > * Primer: I can see treatment of "OWL (2) Full", but I'm not yet > familiar with the Primer, and we still have some time for it, so lets > defer the treatment to later times. > > * No other document talks about OWL (2) Full. > > > Estimated Effort: > ----------------- > > All changes are technically easy to perform (renamings, and a little bit > of explanation in the RDF-Based Semantics). I think that the changes to > the RDF-Based Semantics will take me about 2 hours. The effort for > Conformance will take hardly 15 minutes. Even less for NF&R, if any. So > there will not be any risk for our publication schedule. > > > References > ---------- > > [1] > <http://www.w3.org/2007/OWL/wiki/index.php?title=Conformance_and_Test_Cases&oldid=21446#Syntactic_Conformance> > [2] > <http://lists.w3.org/Archives/Public/public-owl-comments/2009Mar/0077.html> > [3] > <http://www.w3.org/2007/OWL/wiki/index.php?title=Document_Overview&oldid=21428#Semantics> > [4] > <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&oldid=21711#topic-ont-ontology> > [5] > <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&oldid=21711#Vocabulary_Terms> > [6] > <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&oldid=21711#topic-ont-ontology> > [7] > <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&oldid=21711#topic-intro-ontologies> > [8] > <http://www.w3.org/2007/OWL/wiki/index.php?title=RDF-Based_Semantics&oldid=21711#Appendix:_Changes_from_the_OWL_1_RDF-Compatible_Semantics_.28Informative.29> > > > Please let me know (soon!) what you think of this proposal! > > Regards, > Michael
Received on Thursday, 9 April 2009 17:02:38 UTC