[Ietf-languages] Proto-languages

[Richard Wordingham <[e-mail]>]

Greetings,

Some while ago, we had a brief discussion of the English Wiktionary's
non-conformant usage of language tags.  I have since discovered that
these non-conformant tags were being included used as language tags in
the HTML of the web site!  However, sentiment amongst the editors has
moved towards it being a good idea to emit conformant language tags, to
the extent that we are talking about generating a table converting
Wiktionary language tags to BCP 47-conformant tags, not necessarily
preserving Wiktionary distinctions.

There are roughly 624 non-compliant Wiktionary language tags, whereas
there are only about 520 private use language subtags available.  A
programme to eliminate the need for private-use subtags will result in
a flow of registration requests.

A lot of these non-compliant language tags are for proto-languages, and
I foresee major problems in registering.  ISO 639-3 prohibits
reconstructed proto-languages, and the number of documents produced in
proto-languages is quite low.  Is there any point in trying to register
them with ISO 639? If so, how do we do it?

Another possibility I see is to register a variant subtag 'proto'
prefixable by any language subtag denoting a language family (formally,
we might have to generalise to anything with the scope of 'collection')
to denote the proto-language of the family.  If ISO 639 automatically
rejects mere proto-languages, is this in principle acceptable?  We
might have to list individual prefixes for the variant subtag, as
roa-proto would denote some variant of Latin (cf. la-peano for
Interlingua).

A third possibility is to go for BCP 47 5-8 character language subtags.
 Would formal ISO 639 rejections be required?

Richard.