Flavors of Hepburn (was: Re: Ietf-languages Digest, Vol 81, Issue 41)
randy_presuhn at mindspring.com
Mon Sep 28 03:00:05 CEST 2009
> From: "Doug Ewell" <doug at ewellic.org>
> To: <ietf-languages at iana.org>
> Sent: Sunday, September 27, 2009 10:16 AM
> Subject: Flavors of Hepburn (was: Re: Ietf-languages Digest, Vol 81, Issue 41)
> I would be surprised if anyone familiar with romanizations of Japanese,
> or any other Asian language, would expect a variant subtag referring to
> a particular romanization to apply only if the rules are followed
> precisely. That would be like saying that "en-US" can apply only to
> "proper" American English as described in an authoritative,
> prescriptivist dictionary and not to any regional varieties, nor to
> non-standard grammar or spelling (such as "alright" :).
> This is one reason I didn't want "ISO 3602" or similar to appear in the
> Description field. Anyone familiar with these romanizations should be
> able to tell what is intended by "Hepburn romanization" or "Kunrei-shiki
> romanization," and should not be thrown off by additional specificity
> that might imply only perfectly transcribed text fits the subtag.
I would strongly object to enshrining this (assuming fuzziness) as a
general principle, regardless of what happens with this particular request.
Whether a variant is broadly or narrowly defined should depend on the
needs of the applicant. If the applicant requests and needs a precisely
defined variant, then that is what should be registered. If what is
needed is something broader or more fuzzy, then the registration
should reflect that need. From the discussion of Pinyin and other
Chinese romanizations, along with what we've learned about the
situation with Japanese romanizations, I would hope that it is
abundantly clear that the situations in different languages vary
> I think it is clear by now that only 'hepburn' and 'heploc' should be
> registered at this time, because some participants (not all) feel we
> should only register subtags for which there is a clear and present
> user-expressed need, and because Mark has expressed this concern over
> the continuum of romanizations. We can always return to Kunrei-shiki
> and Nihon-shiki in the future, maybe the very near future, but debating
> them should not further delay registration of Frank's Hepburn subtags.
Agreed. I think the discussion also reveals that attempting to construct
a Romanjistammbaum, while perhaps an interesting academic exercise, would
probably be of little practical value in dealing with actual transcribed data,
and may actually be futile given the variation seen "in the wild".
> I do also think we should try to resolve the schism on this list between
> those who want to "tile the plane" and those who want to register only
> what the user's project needs right now. We seem to get caught up in it
> every time we get a request, and it frustrates both list members and the
I am very much in the "register only what is asked for" camp, but...
I think it is helpful to get some idea of what the plane looks like,
in order to make sure the registrant's intent is understood and to make
sure the description / references are sufficiently precise to cover
the varieties actually intended.
More information about the Ietf-languages