ACTION-430, Move specialization sections to appendices

At the last telecon I  was tasked to explain why ACTION-430,
http://www.w3.org/2005/rules/wg/track/actions/430
should be scrapped and the following resolution reconsidered:

    RESOLVED: make "specialization of FLD" sections (of BLD) appendices, leaving standalone sections in place, and making both standalone and specialization normative.


1. I was surprised to actually see this as a resolution. (I missed it when
   reading the minutes).
   There was no vote on this proposal -- only a discussion. During the
   discussion I only said that I'll think how to best address CSMAs
   concerns.

2. The conditions that existed when we were discussing this issue do not
   exist any more.

   The draft we discussed was a first reasonable draft after a COMPLETE
   rewrite of the BLD (of the October document).  After that I addressed
   over 60 major and medium-grade comments by Stella, Igor, Leora, Jos,
   Harold, and others (not counting minor things). As a result, several
   sections had to be moved, merged, raised level, etc. -- all without any
   resolutions. I spent enormous amount of time thinking about the
   structure of the document and implementing changes (several full days).
   
   Insisting on sticking to a resolution, which was not properly voted on
   and whose premises do not really exist any more is not proper.

3. None of the three formal reviewers of the draft requested this change
   and one (Igor) explicitly said that he prefers the dual way BLD was
   presented.

4. As I said, the new documents are the result of serious thinking about the
   grand schema of things. I think all logic (and later non-logic also)
   dialects should be presented as a specialization of FLD or of a similar
   framework. FLD drastically lowers the bar for the introduction of new
   dialects, and it is easy to envision that some dialects will be specified
   *only* as specializations of BLD. For instance, an LP dialects based of
   the well-founded semantics or stable models does not need direct
   specification because their audience is sufficiently sophisticated in
   various logical approaches.

   The BLD specialization from FLD is thus more important for the grand
   schema of things because it shows, by example, how other dialects can be
   defined. Delegating this to an appendix blurs this important message.
   This will also lead to great variance between the specifications of
   different dialects. Some will place the specialization part in the
   appendix, some will have only the specialization part, and some will not
   bother to include it at all, thereby breaking the RIF framework.
   


	--michael

Received on Sunday, 30 March 2008 03:24:52 UTC