RFC3066bis: looking ahead

Peter Constable petercon at microsoft.com
Thu Jan 22 19:49:29 CET 2004

> From: Debbie Garside [mailto:debbie at ictmarketing.co.uk]

> If you read my message again... you will note I made no reference to
> 639-6

True, but an industry standard will not make normative reference to
something that isn't an industry standard. The options for your alpha-4
identifiers to be used in RFC 3066 or a successor are ISO 639-6 or
individual registrations on a case-by-case basis. I didn't think you
meant the latter.

> As to implementation issues... I would welcome your thoughts and
comments on
> an alpha4 system...

As I said earlier, it is premature to enumerate all the implementation
issues that might need to be considered. I'd add that if you want to be
a proponent of a new solution, then you should first become thoroughly
acquainted with whatever problems might exist that are awaiting a
solution and the application contexts in which any solutions would be
used; a proponent should be able to enumerate themself at least some of
the implementation issues so that others involved can see that the
proponent has been taking into consideration issues important to other
stakeholders as she or he devises their proposed solution.

I will, however, offer some things to consider:

What user need within the communities that use RFC 3066 is being met?
What other ways to meet such a need might exist, and is any of them a
better way to meet the need? 

How do the kinds of entities represented by these alpha-4 symbols relate
to the kinds of entities represented by the symbols in ISO 639-1/-2/-3,
and by the complete tags defined by RFC 3066? Is there a coherent
conceptual model for integrating the former with the latter in a
systematic way?

Given (I assume) an additional layer of granularity introduced by
25,000+ entities that introduce one-to-many mappings that likely affect
the majority of entities already provided in IANA registrations or the
parts 1 - 3 of ISO 639, what kinds of usage guidelines should exist?
When is it appropriate to use which?

Is there an appropriate way to integrate these alpha-4 symbols into the
tagging scheme defined by RFC 3066 or successors? If so, how should it
be done?

What are the implications of such integration for pre-existing tags and
their usage? How or to what extent can backward compatibility be

What are the implications of such integration for the use of ISO 15924
alpha-4 identifiers within the same scheme?

How will existing implementations of RFC 3066 (e.g. HTTP) be affected?
Are changes required? What are the implications if there are deployed
implementations that do not get updated soon or perhaps ever?

Peter Constable
Globalization Infrastructure and Font Technologies
Microsoft Windows Division

More information about the Ietf-languages mailing list