ISO meeting in Atlanta, day 2
Posted in Technology on 2005-11-14 00:35
Day 2 started with a presentation by Naito-san about a proposal from him and Komachi-san about a standard format for publishing PSIs. They suggest using Dublin Core to describe the PSIs. They also suggest creating topic maps to describe the PSIs. Finally, they brought up the question of which standards body should do the work (OASIS or ISO). Again, I'll link to the slides once I have them.
This was discussed back and forth a bit, and eventually it was decided to do work within ISO to define how to publish PSIs. The issue of whether or not to define any new formats (or Topic Maps vocabularies) for PSIs were left to be decided as part of this work, rather than up front.
TMDM, clause 7.5
The discussion then moved on to TMDM, where much time and energy was again spent on clause 7.5, with a compromise eventually being sketched out, where the PSIs would move into an annex, get a note disclaiming them having any processing semantics, and a part of the URI would perform the same function. In case the last point is unclear, it was discussed making the URI have the form http://psi.topicmaps.com/keyword/foo or http://psi.topicmaps.com/concept/foo, or something similar. It turned out to be difficult to find a phrase that would work for everyone, so this discussion was eventually tabled, to be continued later.
XTM was then discussed, with one comment from Japan being that the name should change. After a long and heated discussion it eventually stayed unchanged.
Steve Pepper then suggested that the namespace URI of XTM 1.1 should be changed from that of XTM 1.0 (http://www.topicmaps.org/xtm/1.0/) to a URI like http://psi.topicmaps.com/xtm/. This was vigorously opposed by both editors of XTM, but after several hours of discussion they were defeated by a vote in the committee where they were the only ones voting against the change.
Several reasons were put forward for the change, but the one that eventually swayed the committee (minus said editors) was that the domain topicmaps.org is controlled by Michel Biezunski personally, whereas topicmaps.com is owned by the committee. (Formally, Standards Council Canada, on behalf of the committee.) It was felt to be inappropriate for the committee to be using a namespace it did not own, and so this was the reason for making the change.
The editors opposed the change because they argued that this meant changing all the element type names (since the namespace URI part of the names changed, even if the local part stayed the same). This would mean that formally, in XML, the element type names in XTM 1.0 and XTM 1.1 would have no relationship with each other. Effectively, we would have xtm1.0:topic and xtm1.1:topic, which would be different. The editors then suggested that we might as well go the whole hog and fix all the mistakes in XTM 1.0 (which they advanced as a reason for not making the change). The committee jumped at this, and promptly voted through another set of changes to the vocabulary, namely:
- member would now become role,
- roleSpec would now become instanceOf,
- instanceOf would now be required in baseName, occurrence, association, and role,
- parameter would now become scope, and
- the order of instanceOf and subjectIdentity is switched within topic.
So this is the news. XTM 1.1 will be in a different namespace, and will not be backwards compatible with XTM 1.0. Also, the XTM 1.1 specification no longer defines what XTM 1.0 documents mean, so what an XTM 1.0 document now means is anybody's guess.
Dmitry and Graham presented the current state of TMCL, which had a rather strange reception. On the one hand it was felt to be quite good and nearly finished, but the explanation of it in the document was felt to be too complicated. The result, after much discussion, was a breakout session where one team came up with a different way to define TMCL (basically by representing the schemas as topic maps, although there is more to it that I won't go into here), while another team wrote up the formal NPs for CTM and GTM.
A number of minor issues in TMCL were also discussed, but it's getting late, and I won't try to remember what they were.
This is an unofficial report from the winter meeting of ISO JTC1 SC34
Read | 2005-11-12 19:30
Many people have asked what the changes between versions 1.0 and 2.0 of XTM are, and what the rationales for the various changes are
Read | 2006-12-16 19:17