Jon Hanna jon at spin.ie
Mon Jun 16 14:11:19 CEST 2003

So, maybe a
> discussion
> of whether the tag should imply a full Oxford University Press
> style (as far
> as possible within the limits of the technology) would be appropriate.

It wouldn't IMO for a few reasons.

Firstly, a lot of the what could be considered OUP style is not concerned
with the matters that are with scope of RFC 3066, this is not the place to
decide on them. In fact we can't, any hacker worth his salt would assume
that everything in the registration outside of the scope of RFC 3066 should
be ignored.

Secondly, adding such a burden on the tag reduces its usefulness. One of the
reasons given for registering such a tag is that it is the English
orthography used by many documents produced by international organisations.
Such groups tend to also have formalised standards on matters that would be
covered by the OUP style, and will not be in agreement with them (in
particular ISO 8601 dates tend to be used by such groups for the same
reasons that they are used in many technical contexts in counties, including
the US and all EU countries, where other date conventions are more common in
other contexts, this is not OUP style IIRC).

Thirdly the caveat "as far as possible within the limits of the technology"
pretty much wipes out everything beyond language information anyway. The
limits of the technology will vary considerably depending on the technology.
>From where we sit we'd have to assume that the only thing the technology is
capable of doing is at least simple matching based on language tags and at
most that with the addition a few of the simpler machine-human conversions
based on language (it would know whether to call this month June, Meitheamh
or Juin, but anything else it deduces about how to express dates would be
guess-work at best unless it is informed by another source). Anybody
requiring more for their application will look elsewhere for it.

Fourthly, it would make it far more difficult to deal with any variations in
these matters, whether due to a change in practice or identification of a
mistake in the specification. For instance if we switched from a personal
name convention of "surname, firstname" to "firstname surname" or perhaps
adapting either of those to better cope with Japanese names, then we would
need a new language tag even if nothing else changed.

Identifying just were to split a larger problem into components isn't always
without controversy, but I don't think there's a hacker alive who would say
that you shouldn't split just about any complicated system into components.
en-oxford (whichever variant thereof) would allow you to construct exactly
the sort of all-encompassing style-guide you describe, but in not providing
it in one single stroke it increases both the flexibility of your code, and
also the likelihood of en-oxford being used in a wider context that would
lead to your style-guide interoperating with other code.

More information about the Ietf-languages mailing list