RE: [xtm-wg] Status of Core Deliverables document

[Steve Pepper <pepper@ontopia.net>]

[If you are not interested in process issues related to the development
of the XTM spec, feel free to ignore this posting.]

I wrote:
> > One issue that will have to be decided very early on at the
> > Paris meeting is the status of the 1.0 Core Deliverables
> > document.

At 13:13 15/01/01 +0100, Michel Biezunski replied:
>The only changes in the DTD that are compatible with the above statement are
>those which do not invalidate any document instances conforming to the DTD
>published in December.
>...
>But it would be better to avoid even these kind of changes if we can. We all
>need to make a special effort to consider that at one point we have a spec
>out, and that it reflects the current state of consensus. I don't think
>anybody can ever say that a standard is definitive, but at least we have
>to make it clear that investments which are made by users on topic mapping
>their information have to be preserved in all cases. This has already
>started since 1.0 is out. If we don't take this into account, nobody will
>want to take the risk of investing in a ever moving standard, which is not
>properly maintained.

I understand Michel's concerns and I share them: The market needs a good,
stable, XML-based topic map standard as soon as possible. But not at
any cost: We must achieve the right balance between "good", "stable",
and "as soon as possible".

There are several factors that could lead to a perception of instability.
One would be if the spec were to change repeatedly (Michel's "ever
moving"); another would be if the process by which the spec has been
created were not to inspire confidence.

I am worried that the publication of the Core Deliverables document might
be in breach of TopicMaps.Org's charter, which states:

    5.1.2. XTM Approval
    Formal adoption and all revisions to XTM shall require a two-thirds
    majority vote of the Participating Members.

The participating members as a whole were NOT given the opportunity to
vote on this document. The only vote that took place was one in Dallas
which authorized the editors to complete the spec based on the
consensus that seemed to have been achieved on Sunday Nov. 12th between
the modelling group and the syntax group.

Despite heroic efforts, the editing team (the Gang of Four: S&M&M&S)
were not able to complete the spec. Not only that, they also discovered
what they felt were a number of serious bugs in the results of the
Dallas meeting, and made extensive revisions to the DTD accordingly.
(In fact, the belief that such bugs existed and the desire to iron them
out through maraton con calls was probably the reason why they were not
able to complete the spec.)

Unfortunately we still only have incomplete minutes from the Dallas
meeting, and they do not record *any* of the many formal decisions
that were taken regarding the syntax. (For example, a vote was taken
on whether or not to have <baseName> as a subelement of <occurrence>.
The decision was no, but all the same it's there in the version of
the DTD in the Core Deliverables. There are a number of such examples.)

It is my feeling that the revisions made between Nov 12 and Dec 4
(the date of the Core Deliverables document) are so extensive and
so little discussed in the AG as a whole, that they go beyond the
mandate that was given in Dallas.

Several comments on this mailing list during the last couple of
weeks suggest that the syntax no longer accurately reflects the model.
If that really is the case, then we absolutely must fix the problems,
even if that means changing the Dec 4 DTD in ways that are non-
backwards compatible.

Given that that document has not been formally approved, this really
ought not be a problem.

Steve

--
Steve Pepper, Chief Technology Officer <pepper@ontopia.net>
Convenor, ISO/IEC JTC 1/SC 34/WG 3 (Information Association)
Ontopia AS, Maridalsveien 99B, N-0461 Oslo, Norway
http://www.ontopia.net/  phone://+47 90827246/


To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com