ISO meeting in Seoul, day 1 | Larsblog

View from the meeting room (Gangnam Tower)

The spring ISO meeting is this year in Seoul, and this is actually the first time in the history of SC34 that it has a meeting in Asia. This was more than a little overdue, given that there has been heavy Japanese attendance at these meetings for more than a decade, and heavy Korean attendance for the last 4-5 years. Of course, I personally am also very happy to get a trip to Seoul, which is also a very interesting city. (More on this later, if possible.)

TMQL

The first day of the Seoul meeting was dedicated to TMQL from early morning until we finished at 1800. It consisted mostly of Robert Barta presenting his tutorial of the new draft, interspersed with discussions and questions about various points in the draft. There was quite a bit of feedback, but the general impression was that the draft has made big steps forward.

Robert in action

Most of the discussion related to minor points in the semantics (how are ID references to topics defined?, what is the proper semantics for path expressions in SELECT and ORDER BY?, etc), syntax issues (proper syntax for AND, OR, ...), and so on. The presentation of the language (with discussion) took the entire day, so there was really no time to make any definite decisions.

Topic Maps metamodel

One thing that TMQL does is to assume that certain predefined topics exist with specific semantics. One example of this is what the tutorial calls xtm:subject, which (in TMQL) is the supertype of all typing topics and the type of all topics. Robert expected this to exist in TMDM, and was disappointed to find that it did not. This is what inspired me to write the previous posting during Robert‘s presentation of TMQL.

I presented the previous blog posting at the end of the meeting, and met with general agreement. This raised the question of where it would go, and how it would be expressed. We came to the conclusion that since this is a formal definition of the TMDM semantics it seemed appropriate to make this a part of the mapping from TMDM to TMRM. This also solved how to express the mapping, since we can then use TMRM path expressions. (tolog is obviously not appropriate, nor is TMQL, since this is supposed to be the basis for TMQL.)

My impression is that there are some things in the previous blog posting that are lacking from what Robert needs in TMQL, but we can bridge that gap later.

Schedule

At the end, Steve took over to set up a schedule for the delivery of the next drafts, and the votes on these. I don‘t have it to hand, so I can‘t reproduce it here, but I‘ll either post it or a link to it some time during the week.