[sc34wg3] Comments on XTM 2.0

Steve Pepper sc34wg3@isotopicmaps.org
Wed, 11 Jan 2006 02:30:43 +0100

* Lars Marius Garshol
| The editors have now produced new drafts of TMDM and XTM 1.1 at
|    http://www.isotopicmaps.org/sam/
| ...
| XTM will go to review (still as FCD), so that it can get a proper  
| review before it is finally nailed.

The plan is to go to FDIS ballot on 13250 Part 3 early enough that
we have the result in time for the next meeting of WG3, which takes
place in Seoul, Korea on May 29-June 1 2006.

This means that we have until the end of February to make comments
on the XTM draft and discuss any issues arising.

I urge everyone to give http://www.isotopicmaps.org/sam/sam-xtm/ a
thorough review.

Here are my comments:

1. Version number should be 2.0

Now that we've decided to break backwards compatibility and make
some quite radical changes, we should up the version to 2.0, not
1.1. This would reflect that the changes are substantial. It would
also, I think, better convey our intent that the schema remain
stable from now on. "1.1" implies that there could well be a 1.2
or a 2.0, whereas "2.0" is often associated with the first release
of the *real* version of a product (after the bugs have been
ironed out of 1.0, which really is the case here).

This is a marketing argument, of course. However, I cannot think of
any technical arguments why we should only go to 1.1. Are there

2. GI for names should be <name>, not <topicName>

I see no reason to use a longer GI when a shorter one will do. We
talk about "names, occurrences, and associations", so why would we
choose <topicName> to go with <occurrence> and <association>?

Perhaps because that is the name of the corresponding information
item in TMDM? If that's the reason, the same logic should apply to
<role>, which should then be changed to <associationRole>.

But I don't think this is necessary: The <role> element is a
direct subelement of <association>, so its meaning is clear enough.
The same applies to <name>, which is a direct subelement of

Shorter is better.

3. Error in DTD (content model of topicName)

  <!ELEMENT topicName
     ( itemIdentity*, instanceOf?, scope?, value, variant* )

instanceOf should be required, not optional, as it is elsewhere
(including the normative RELAX-NG schema).

Otherwise, I think XTM 2.0 looks very much better than 1.0. My
congratulations to the editors!


Steve Pepper <pepper@ontopia.net>
Chief Strategy Officer, Ontopia
Convenor, ISO/IEC JTC 1/SC 34/WG 3
Editor, XTM (XML Topic Maps 1.0)