RE: [Ltru] Re: Is 639-3 bogus ?

Can I suggest that, next time, it would help if you stated clearly at the outset what your concerns are. It's taken me several rounds going back and forth with you to drag out what it is you're going on about.

Re entries in the 639-3 draft code table for ancient and constructed languages, the Linguist List was the source for those, and presumably they had a reason why each of these was deemed to have a need. If you have a question about the usefulness of orq or any other entry, you can contact Anthony Aristar or Debbie Anderson.

Re the characters used in the names, the source for the large inventory of entries of modern and recently-extinct languages came from a database that has a history going back several decades. By the time 639-3 is published, some of the spellings may be revised, or some may be revised over the life of maintenance for the code set. It was not considered a prerequisite to normalize the spellings to conform to any particular expectations prior to initial publication.

If you see two entries -- two distinct IDs -- then the language varieties denoted are deemed to be distinct languages. The fact that they have similar spellings is coincidental. The language names connected with "byj" and "bmn" are even more similar than the pair you cited; that doesn't mean that the entries are suspect.

The source for the majority of new entries in 639-3 has been published for many years -- it's in its fifteenth edition -- and has been widely used and reviewed by linguists throughout the world. 

IMO, there's no particular reason for members of this WG to be second guessing whether the content of ISO 639-3 is good enough for use in the LSR.

Going back to where this started, you responded to comments I made about how the type information in 639-3 may be useful to a user. "That's the job of a comment or description," you said. It's still not clear to me if you had a comment to make wrt John's proposal or not.

Peter

-----Original Message-----
From: Frank Ellermann [mailto:nobody@xyzzy.claranet.de] 
Sent: Tuesday, October 10, 2006 2:59 PM
To: ltru@lists.ietf.org
Subject: [Ltru] Re: Is 639-3 bogus ?

Peter Constable wrote:

> Is it useful to know that orq is an individual, constructed language
> and that that language is Orcish? Yes.

alpha-3 is a finite set, if any language of 30 words gets its own code
639-3 is in trouble, and the subtag registry with it.

> Again, what is the point of these questions?

Protecting the subtag registry from bogus entries while it's on a one
way street (nothing ever removed) with a known end (26*26*26).

> And explain to me how this is not a random attack on 639-3, since
> that is certainly how it appears to me.

Of course it's a random attack, I looked into the source, because John
proposed to preserve its Language-Type info, and the "C" attracted my
attention.  There are not many "C", and "orq" is one, and as it happens
I thought (yesterday) I know what "orq" is about.  And today I'm sure.

I can't judge anything else, e.g. why "bet" offers Béte and "bev" Bété,
it could be a typo, it could as well be completely unrelated languages.

The source for the draft is Latin-1 (another detail I didn't know, and
I checked that it's really Latin-1 and not windows-1252).  In some
cases like "aue" =/Kx'au//'ein or "gnk" //Gana I'm not sure if that's
as it should be, it appears to be odd.  Or maybe "//" has a well-known
meaning, and I just don't know what it is.

> we can revise 2.2.1 to say whatever we think is needed.

What the quoted promise already says just adding ISO 639-3 would do it.

Frank

_______________________________________________
Ltru mailing list
Ltru@ietf.org
https://www1.ietf.org/mailman/listinfo/ltru

_______________________________________________
Ltru mailing list
Ltru@ietf.org
https://www1.ietf.org/mailman/listinfo/ltru