W3C home > Mailing lists > Public > www-webont-wg@w3.org > June 2003

Response to QA comments, Comment on QA draft

From: Sandro Hawke <sandro@w3.org>
Date: Wed, 11 Jun 2003 23:14:05 -0400
Message-Id: <200306120314.h5C3E55m014202@roke.hawke.org>
To: www-webont-wg@w3.org

I've reviewed the materials.  I don't see any big problems.  Their
review [1] [2] suffers mostly from treating S&AS as if it were a
standalone specification; Dan Connolly already pointed that out to
them in his preliminary reply [3].

In Jeremy Carroll's comments [4] on their draft [5], I would include
an apology for sending this so late, but then I clearly don't
subscribe to the included advice ("never apologize, never explain").
I'd also include a link to http://www.w3.org/2001/01/mp23 .

Dan left a few open points for WG discussion.  Here they are with my
recommendation of what we should do about them, offered to give us at
least a baseline for discussion.

  Guideline 9. Allow extensions or NOT! [6]

      They want us to say whether extensions to OWL are allowed or
      not.  I think the answer is they are not.  If you add anything
      to DL it's no longer DL.  But given RDF's same-syntax semantic
      extension form, one could also say extensions are of course
      allowed; they just don't have any semantics in OWL.  Are we
      concerned about vendors touting software which implements
      OWL-plus-proprietary-extensions, and the interoperability
      problems that might create?

      I propose adding at the end of TEST Section 4.2.1 (Syntax
      Checker) something like:

           OWL is not an extensible language.  The underlying RDF
           graph may contain non-OWL terms (subject to the
           restrictions in 4.1.1 on OWL Lite and OWL DL) which are
           used as RDF extensions, but they do not extend the syntax
           or semantics of OWL itself.
  
      or just

           OWL is not an extensible language.

  Guideline 13.2    Distinguish normative and informative text.  [8]

      They want us to indicate each section as being "Informative"
      or "Normative", not just indicate the informative sections in a
      normative document and leave the rest unlabeled as we now do in
      S&AS. 

      I suggest we label the five top-level sections in S&AS
      "(Normative)".   (Or maybe the Introduction is only
      Informative....) 

  Guideline 14.  Provide test assertions.   [9] 

      It looks to me like the mapping they want between S&AS and the
      test cases would be very difficult to construct, unless we have
      it somewhere I haven't noticed, and not of tremendous value.  I
      suggest we respond that we find their rationale insufficiently
      motivating given we already have an extensive test suite.

I think the rest of their issues are addressed by Jeremy's message [4]
(to be sent) and Dan's message [3] (sent to Karl).  They requested
that discussion of their comments to us [2] and feedback (like
Jeremy's) on their drafts go to www-qa@w3.org.  While we could formaly
separate our comments on their documents (which are months past the
Last Call deadline) from our response to their comments on our
documents, I think it's simpler at this point to combine them.  I
suggest we either:

      1.  send them and Karl our response to the above Guidelines (and
          ask if they find the response satisfactory, etc, etc) with
          pointers to Jeremy's and Dan's messages in the archives, or 

      2.  send a more thorough reply, incorporating the points made in
	  Jeremy's and Dan's messages, rather than just referring to
	  them in their present archival form.

This subject is on the Agenda both under Action Item Review (ACTION:
Sandro Hawke, Evan Wallace - review Jeremy's review of the QA last
call so that it can be sent back as a WG review) and "LC comments
status" (DANC: QA Review of owl-semantics (Wed, Apr 30 2003) [no reply
received] - Connolly started discussion, not yet done.).   If nothing
above is contentious, perhaps it can be quickly approved in one of
those slots.

    -- sandro

[1] http://lists.w3.org/Archives/Public/public-webont-comments/2003Apr/0064
[2] http://www.w3.org/QA/2003/04/QA-Rev-owl-semantics-all
[3] http://lists.w3.org/Archives/Public/public-webont-comments/2003May/0002
[4] http://lists.w3.org/Archives/Public/www-webont-wg/2003Apr/0158
[5] http://www.w3.org/TR/2003/WD-qaframe-spec-20030210/
[6] http://www.w3.org/TR/qaframe-spec/#Ck-extensions-disallowed
[7] http://www.w3.org/TR/owl-test/#conformance
[8] http://www.w3.org/TR/qaframe-spec/#Ck-normative-informative
[9] http://www.w3.org/TR/qaframe-spec/#Ck-markup-assertions
Received on Wednesday, 11 June 2003 23:14:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:01 GMT