The meeting was held at 2002-05-24, 1600-1800, in the Victoria room of the Princesa Sofia Intercontinental hotel in Barcelona.
| Thomas Bandholtz | Member | Present |
|---|---|---|
| Lars Marius Garshol | Member | Present |
| Steve Pepper | Member | Present |
| Mary Nishikawa | Member | Present |
| Motomu Naito | Member | Present |
| Bernard Vatant | Member | Present |
| Peter Pappamikail | Prospective member | Present |
| Patrick Durusau | Visitor | Present |
The proposed agenda was unanimously approved, as were the minutes of the previous meeting.
Peter Pappamikail expressed interest in becoming a member, and so became a prospective member.
Thomas Bandholtz reported that the OpenGIS Consortium is working on a specification for gazetteers, which might be an interesting user of the work done by the GeoLang TC.
Patrick Durusau reported that he had contacted the Ethnologue project, which is now intending to become the code set for languages. Patrick will establish contact between the chair and the Ethnologue project lead: Gary Simons.
The chair expressed the opinion that until a working draft had been published continuing liaison work would be of little use, as it would be difficult for potential liaisons to know what GeoLang had to offer them. Once a draft has been published, however, this will change, as there will be an example for them to look at. This view was generally accepted.
This issue was discussed at length, but in the end it was decided to define two published subjects as given below:
This resolution was voted on at the meeting, receiving only yes votes, except for Bernard Vatant's "abstain" vote.
At this point in the meeting Peter Pappamikail and Thomas Bandholtz departed, having to catch their respective flights home.
The resolution proposed in the agenda was adopted without much debate, as it corresponded well with the policy arrived at in the PubSubj TC on the morning of the 24th.
Resolution: Make an HTML page entry page for each PSI set, let that page contain informal metadata as well as definition text for each published subject, and a link to an XTM file. The XTM file will be similar to those already proposed by Murray Altheim and Lars Marius Garshol.
This issue provoked much debate, but was in the end resolved to general satisfaction. One published subject identifier will be defined per code, but only one subject per language/country. That is, there will be multiple URIs for each subject, but only one of these will be recommended. The recommended URI should be the one based on the numeric codes, as these are stable. A normative mapping between the codes should be provided in the form of an XTM topic map.
It was decided that the XTM file should provide only the English and French names given in the standards, appropriately scoped. There should be no instanceOf references to the classes, no plain-text subject indicator, nor should the codes from the standards be repeated.
It was decided that as we recommend the use of the numeric codes as the only stable codes we are in any case implicitly breaking with the old URIs given in language.xtm. For this reason it was decided to not use the old PSIs, but to create entirely new ones. It was decided that a normative mapping to the old PSIs should be provided, however.
After some discussion, it was decided that the preferred URI forms (in order of preference) were:
The chair was instructed to do the following:
The next meeting will be held following the Extreme Markup Conference in Montréal, August 10, 2002.