Generic variants and Armenian dialects

[Frank Ellermann]

Doug Ewell wrote:

> we have implemented "spelling reform" variant subtags for
> only one language, and the practical need for those has been
> called into question

The exact variant used to be relevant in some situations like
schools.  There was a transitional period with both variants
"allowed" (in DE, I can't tell you more about AT, CH, etc.),
that ended some weeks ago with a final version.  The final
version isn't exactly the same as "1996", it's more relaxed.

For almost anything published today I'd expect it to follow
that final version - more or less, e.g. more for schoolbooks,
less for newspapers.  In say twenty years "1996" could be
marked as deprecated, because it's simply redundant.  Or the
the mainstream is then already so different from "1996" that
keeping it makes sense.  For "1901", just leave it alone, it
can be used with spellcheckers, e.g. for new editions of old
(pre-1996) books.

Finding other tags for similar cases if necessary should be
easy.  But something like "neuerecht" or "newspell" has its
own drawbacks, neu / new is relative, and available only
once per prefix:  The 1901-variant was new a century ago.

Maybe folks find other uses for "1996", it could indicate
spellchecked.  A probably bad analogy:  1901 ~ frameset,
1996 ~ strict, no embellishment ~ transitional.

Back to the topic of generic and Armenian variants:  I'm
still not convinced that "eastern" in any shape or form is
a good idea, if it in essence means "non-western".  There
is no "nonboont" in the registry, and I don't miss it.

Frank