Revised request: Japanese transliteration variants

Randy Presuhn randy_presuhn at
Thu Sep 3 02:39:20 CEST 2009

Hi -

> From: "John Cowan" <cowan at>
> To: "Frank Bennett" <biercenator at>
> Cc: <ietf-languages at>
> Sent: Wednesday, September 02, 2009 3:03 PM
> Subject: Re: Revised request: Japanese transliteration variants
> > I am not intending to claim "hepburn" as the unique designation for a
> > particular system.  I only mean to point at the ALA-LC tables as the
> > most complete and accessible set of transliteration instructions for a
> > commonly used Hepburn romanization method.  If another party wishes to
> > register a subvariant, and it became necessary to distinguish this
> > common form from other variants, I would be happy to apply for a
> > subvariant at that point.  But I'm not sure there will be a demand for
> > that, and it seems better to keep the spec as simple as possible to
> > satisfy the current need (which is just to distinguish "text of the
> > hepburn sort" from "text of the kurei sort").
> In that case, reference to "revised" or "modified" should be removed
> from the draft of the entry itself (as opposed to the supporting
> information, where it can remain).

I'm a little puzzled.  I *thought* the request was to register a *specific* one
of the various flavors of Hepburn, leaving open the possibility that others
might someday want to register other flavors of Hepburn.  It's clear to me
how such a thing would be used, and how it would be useful.  It was only
somewhat unclear whether the original request intended "revised" or "modified",
since it used both names.

It now sounds like at least some on this list are interpreting this as a request to register
a variant subtag that would be usable for any of the various flavors of Hepburn.
If someone needed to tag something as being in a specific flavor of Hepburn,
an additional variant subtag would be needed. Though I can understand the
architectural appeal of such an approach, it's less clear to me how such a thing
would actually be useful, except perhaps for auto-tagging of strings that look
the same in multiple flavors of Hepburn.  Earlier examples in this thread
indicated that there were contexts in which a specific Hepburn flavor would
be required, which would lead me to conclude that there is a need to distinguish
at least that particular variety from "generic Hepburn".

If we already know that what is to be tagged is a specific flavor of Hepburn, and
admit the possibility that there will be a need to distinguish the other varieties,
and do not want "legacy" material to be ambiguous in its tagging, then we need
to register that specific flavor now, either as a subtag of ja-Latn or of ja-Latn-hepburn,
whichever approach finds favor here.  Otherwise, we'll be creating a situation
analogous to 'zh' and 'cmn'.


More information about the Ietf-languages mailing list