[Top] [All Lists]

Re: Last Call- Tags for the

1995-01-09 11:42:10
At 1:51 PM 1/8/95, Harald(_dot_)T(_dot_)Alvestrand(_at_)uninett(_dot_)no wrote:
I see two possible problems with using references to an authority
that changes over time:

- Tags used in old data may be left without a resolution mechanism
 (such as when a language or country disappears or is renamed).
 This may be solved by keeping a registry of old names around, and
 should not be a terrible problem.

- Tags used in old data may be turned into valid names for something
 new because of the reassignment of the name. This is a real problem,
 but I think the registry agencies are fairly aware of this.
 The changes made so far seem to be corrections.

The problem that the real world changes is not going to go away soon.
ISO has been fairly consistent in the past about not handing out
new country codes that reuse an old allocation. Other problems
(like Moscow Russian now being ru-RU rather than ru-SU after the breakup
of the Soviet Union) we have to live with.

The references to ISO 639 are intentionally equipped with dates, and
the registry maintainers are listed, exactly to make people aware that
these things change over time.
I don't think we should do anyting different; second-guessing ISO is not
something the IETF should be doing, IMHO.

Thanks for the praise!

       Harald A

Theoretically speaking, a better scheme would be one which tagged languages
without reference to countries. Isn't the RU or SU for Russian sort of
redundant? Russian is a language spoken in many places. Of course, there
are geographic variants of languages, but those could also be described
without reference to the geopolitics-du-jour. Languages come, go, and
change at a slower rate than countries, and for text, tagging it with a
country serves no useful purpose. What you want to identify is the language
and *only* the language.

Of course, in the absence of an existing alternative standard that tags
languages in this way, it is probably beyong the scope of what IETF should
be doing to invent one, so the ISO scheme is probably the best we can do.

David Goldsmith
Senior Scientist
Taligent, Inc.
10201 N. DeAnza Blvd.
Cupertino, CA  95014-2233

<Prev in Thread] Current Thread [Next in Thread>