Suggestion: registration of variant subtags for Aluku, Ndyuka, and Pamaka

Doug Ewell doug at ewellic.org
Mon Jan 26 14:50:33 CET 2009


"Lang Gérard" <gerard dot lang at insee dot fr> wrote:

> The number of reserved tags from qaa to qtz is 26.17= 442, so that 17 
> 576 - 442 = 17 134 possibilities are remaining  available to assign 
> codes to about 7 700 languages

There are 20 letters from 'a' through 't', not 17.

> (I recently concurred to reduce their number, by obtaining from ISO 
> 639-3 the deprecation of the non-existent "europanto (eur)" !)

This doesn't reduce their number, because the MA doesn't reuse code 
values.  "eur" is no longer available for use by any language name. 
This is true for the other 130+ code elements retired by ISO 639-3 since 
its inception.  (ISO 639-3 "retires" code elements, they don't deprecate 
them.)

I also forgot to count 114 code elements used by ISO 639-5 from the same 
3-letter code space, which further reduces the choices of available code 
elements.  So I'm not completely OK with my arithmetic either; I 
underestimated the problem.

> Nevertheless, this result says that the ratio of occupation from 7 700 
> to 150134 is less than 51 /100, that does not very much support  the 
> thesis that a systematic visual association between the reference name 
> of each language and the code element  for the representation of this 
> language name is generally impossible (I do not write that this is 
> always possible, maybe a few benign exceptions remain necessary). On 
> the contrary, my position is that, by making sometimes astute choices, 
> this task can be rendered possible.

Letter frequency distribution in language names -- the fact that 't' and 
'n' and 'r' are used much more frequently that 'q' and 'x' and 'z' --  
would tend to argue against your position.  Most of the good ones are 
already taken.

> So, we [ISO 3166-1/MA] are left with 676 - (43 + 51) = 582 really 
> available code elements, to be compared with 246 active entries, and 
> 51 more not reusable code elements, that gives a ratio of 297/582 = 
> 51/100 that is exactly comparable with the ratio for alpha-3 ISO 639-3 
> code elements !

So you can appreciate the nature of the problem, especially when a new 
country comes along whose name begins with a common letter like 'M' or 
'S', and you have to resort to something like "MF" for "Saint-Martin."

--
Doug Ewell  *  Thornton, Colorado, USA  *  RFC 4645  *  UTN #14
http://www.ewellic.org
http://www1.ietf.org/html.charters/ltru-charter.html
http://www.alvestrand.no/mailman/listinfo/ietf-languages  ˆ



More information about the Ietf-languages mailing list