Multiple preferred values [was: Add retired tag "eml" to the IANA registry]

[Doug Ewell]

"Michael(tm) Smith" <mike at w3 dot org> wrote:

> If it's anticipated that there's any possibility in the future of tags 
> being split, it would seem like there might be a need to have a means 
> to specify multiple preferred values.

Randy said that "BCP 47... doesn't allow splitting a tag," but I think 
what Michael means here, and how many people will interpret the actions 
of ISO 639-3/RA and their effects on the Registry, is that when the RA 
retires code element 'eml' and creates 'egl' and 'rgn', that has the 
practical effect of "splitting" 'eml' into 'egl' and 'rgn'.  That is 
something which we can expect to happen many more times, and which was 
anticipated in the development of RFC 5646.

The Preferred-Value field is provided to allow automatic one-to-one 
mapping.  An example is that, because ISO withdrew 'iw' for Hebrew in 
favor of 'he' some 20 years ago, the Registry entry for 'iw' has a 
Preferred-Value of 'he' to assist matching and automatic retagging.

If you provide multiple Preferred-Values, you really haven't solved the 
problem in the way you might think: should a given piece of text marked 
'eml' match a search for 'egl', or not?  Since there is no reliable way 
to answer this question automatically, a human-readable Comments field 
turns out to be a better solution.  See the region subtag 'CS' to see 
how this sort of multiple-split problem has been handled in the past.

--
Doug Ewell  |  Thornton, Colorado, USA  |  http://www.ewellic.org
RFC 5645, 4645, UTN #14  |  ietf-languages @ http://is.gd/2kf0s ­