Re: [geolang-comment] Quo vadis, GeoLang?

[Lars Marius Garshol <larsga@garshol.priv.no>]

* Lars Marius Garshol
|
| Since then, however, a number of people now about to join the TC
| have suggested other code sets that should also be covered. The
| point of this discussion is basically to reach agreement on what we
| want to do. Personally I believe that we should let history be
| history and do whatever the community has the most use for.

* John Cowan
| 
| I agree, and I propose that we cover the following codesets:
| [...]

Then we seem to be in agreement on the our basic goals, at least.
I think we are bound by history in one sense, though: the ISO 639 and
3166 code sets are already out there, and in need of correction, so I
think we must do these two before we do any of the other code sets.

I expect that to be acceptable to everyone involved.

| RFC 3066 language codes (which are a superset/refinement of
| ISO 639-2, and are now the standard for xml:lang attributes).

I think covering RFC 3066 makes perfect sense as an addendum to the
ISO 639. Users can then decide which codes they wish to use, and third
parties can publish different kinds of mappings between the two code
sets. (One mapping might assert that #zh and all the #zh-* codes refer
to the same language, while another might use associations to indicate
the relationships between then.)

Having covered ISO 639 it should be really easy to cover RFC 3066, and
I think the value of RFC 3066 is great enough to justify doing it.

| This is an IETF/IANA facility; although these are not technically
| standards bodies, they operate as such de facto, maintaining the
| canonical descriptions of core Internet facilities.

IANA is enough of a standards body for me.
 
| UN/LOCODE, an extension of ISO 3166-2 for cities (particularly port
| cities), airports, and other notable locations.  This is maintained
| by the United Nations.

I wonder if this is the same code set that is documented on these
pages: <URL: http://www.un.org/Depts/unsd/methods/m49.htm >. If it is
we've already been thinking of covering it, since Murray Altheim
already has done an initial proposal on it.

If not I think it sounds like a code set that might be useful, but I
don't really know anything about it. Could you say something about who
uses it?
 
| I estimate the resulting total map size to be on the order of
| 100,000 topics (almost all automatically derived from the underlying
| codesets, so not as frightening as it sounds).

As long as we can do this with scripts (and everything we've done so
far was done that way) I have no problem at all with the size of the
UN/LOCODE code set.

I am personally considering whether it's worth our while to cover ISO
15924 (codes for scripts) and the IANA charset registry (codes for
identifying character encodings). I already have a topic map in which
I need these PSIs, but whether anybody else will actually need this is
something of an open question.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >