No subject
Tue Nov 18 23:43:20 CET 2008
ISO 639-3, which is updated annually, so such discrepancies are to be<br>
expected. The Ethnologue web site (as distinct from the SIL 639-3
web<br>
site) is tied to the Ethnologue printed book, and new editions of that<br>
are issued only every few years. The 1st edition was issued in 1951,<=
br>
and Wikipedia says that the next, or 16th, edition will be issued in 2009.<=
br>
</font></tt><font size=3D2 color=3D#008000 face=3D"sans-serif">JAS: This is=
not
an oversight, nor is it Ethnologue lagging behind the ISO process. A look
back in the 14th ed. of the Ethnologue shows that Akan was a variety there,
as an individual language, with dialects Fante and Twi (among others).
The Ethnologue editorial staff decided for the 15th ed. (online now) that
Akan is the language unit of the level they address, and Fanti and Twi
are dialects. There was lengthy consideration consulting a number of people
(okay, I suspect it was mostly SIL people, but I am reasonably sure they
consulted government representatives and other linguists in Ghana, as well).
There was a strongly unified position that Twi and Fanti are subvarieties
and should not be considered separate languages. I got emails about it
yet again this year, worried that Ethnologue was going to put in Twi and
Fanti separately.</font>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">btw, Ethnologue staff
have made a similar decision with regard to Norwegian (Nynorsk and Bokm=E5l)
for the 16th ed. It is at their discretion. I don't control the Ethnologue's
determination of which ISO 639 code elements they include, or exclude.
I have tried to make a case, though, that the online ethnologue makes it
clear when they are treating a macrolanguage individually (so to speak).</f=
ont>
<br><tt><font size=3D2><br>
> 2. This is more of a request; it would be quite handy if the web<br>
> queries (both for ISO 639-3 and the Ethnologue) also took the BCP
47<br>
> forms, eg.<br>
> <br>
> http://www.sil.org/iso639-3/documentation.asp?id=3Dar<br>
> worked just like:<br>
> http://www.sil.org/iso639-3/documentation.asp?id=3Dara<br>
<br>
I heartily agree.<br>
</font></tt><font size=3D2 color=3D#008000 face=3D"sans-serif">JAS: I will =
discuss
it with my webmaster. I do not think it will be a problem. While ISO 639-3
does not actually contain [ar], neither does it contain [afa], but </font>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">http://www.sil.org/i=
so639-3/documentation.asp?id=3Dafa</font>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">does work, and </fon=
t>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">http://www.sil.org/i=
so639-3/documentation.asp?id=3Dar
does not.</font>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">So I will try to get=
us
consistent.</font>
<br>
<br>
<br>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">And retrieved from M=
ark's
original:</font>
<br><tt><font size=3D2>3. Now that ISO 639-5 is out, are there any plans
to update the<br>
Language Family Index, so as to add the new codes that were missing?<br>
<br>
Eg, so that we'd see on http://www.ethnologue.com/show=5Ffamily.asp?subid=
=3D90017<br>
<br>
Indo-European (ine)<br>
Albanian (sqj)<br>
...<br>
Baltic (bat)<br>
...<br>
Germanic (gem)<br>
...<br>
The table in ISO 639-3 of the language families is interesting (eg ine<br>
: gem : gmw), but (a) incomplete, since it doesn't actually include<br>
the elements of gmw, and (b) not machine readable (grrrh).<br>
<br>
Mark</font></tt>
<br>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">JAS: I think you mean
"the table in ISO 639-5..." not in 639-3.</font>
<br><font size=3D2 color=3D#008000 face=3D"sans-serif">I have a draft mappi=
ng
of Part 3 onto the tree of Part 5, but its (Part 5's) structure is incomple=
te
(as you tacitly point out), and right now we (the JAC) appear somewhat
stalled on how to move ahead. And yes, a machine readable hierarchy is
needed. I know Rebecca Guenther and LoC have been working on that.</font>
<br><tt><font size=3D2><br>
-- <br>
Andrew Watt on Microsoft:
John Cowan<br>
Never in the field of human computing
cowan at ccil.org<br>
has so much been paid by so many
http://www.ccil.org/~cowan<br>
to so few! (pace Winston Churchill)<br>
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F<br>
Ietf-languages mailing list<br>
Ietf-languages at alvestrand.no<br>
http://www.alvestrand.no/mailman/listinfo/ietf-languages<br>
</font></tt>
<br>
--=_alternative 0010745886257509_=--
More information about the Ietf-languages
mailing list