GeoLang TC: todo list

This document presents an overview of the currently open tasks and issues being faced by the GeoLang TC. It is a working document, constantly being updated as tasks/issues are completed/resolved and new ones discovered.

Suggestion for new tasks and issues should be posted to the geolang-comment mailing list.


The liaison-related tasks have been postponed until the TC has some examples of published subjects to show for itself, as this will make the task of communicating with the prospective liaisons much easier.

Unresolved issues

The chair will start threads for the discussion of each of these issues in turn, in numerical order.

2. Names for countries

Should we give both English and French names for the countries in ISO 3166? If so, how do we get hold of the French names?

5. What to use as indicators?

What indicators should we have for the subjects in the ISO 639 and ISO 3166 PSI sets? Do the source standards use anything besides the names as definitions? Do we need anything more? Should the indicators be XTM or HTML? Should we auto-generate a sentence for each subject, using the name, the code, and some notion of subject class?

6. What metadata do we provide?

What metadata should be provided for the two PSI sets?

10. Changes to code sets

How do we handle changes to the underlying code sets? What do we do when a code is removed, changed, or added?

11. Organization of liaisons

What formal organization do we need to set up to ensure that these PSI sets will remain current?

Resolved issues

1. language.xtm PS identifiers

What do we use as the source for PS identifiers for languages? The original PSI set published with XTM 1.0 used the two-letter codes, but not all languages have two-letter codes, so the ones that don't will need to get three-letter codes. This is inconsistent, but avoids breaking references to the old PSIs.

Do we create identifiers based on three-letter codes for those that already have two-letter codes, in addition to their two-letter codes?


3. The "language" topic

Should we create a "language" topic, which can be used as the class for the language instances? If we do, how do we define "language"? If we don't, do we just accept that every user of language.xtm will most likely define their own "language" topic?


4. How many identifiers?

Both ISO 639 and ISO 3166 define more than one code for each subject. Should the PSI sets we create mirror this, or should they only create one identifier for each subject? If so, which one should we choose?


7. What formats to use?

Which formats should we publish our PSI sets in? HTML? XTM? Anything else?


8. The form of the PS identifiers

What form should our PS identifiers take?
Or some other form?


9. What assertions to make about the subjects?

What should we assert about each subject? Names? Should we make them instances of some class? Should we assert the codes as names or internal occurrences, or not at all?


Lars Marius Garshol, 2002-06-02.