On 7 Mar 2014, at 07:06, Juha Hakala <[log in to unmask]> wrote:
> From library point of view, 639-2 is sufficient for the purposes of bibliographic description.
No, it isn’t. I have published half a dozen books which cannot be correctly described by ISO 639-2.
> Selecting the correct language code from 639-3 would be a lot more time consuming and require more often consultation of external experts. Even using 639-2 can be challenging at times; just to give an example Finnish libraries are familiar with, selecting the correct variant of Sami language is not trivial.
Nowadays, however, publishers are often themselves providing the metadata, as with the British Library Cataloguing-in-Publication data. So it is no longer the case that lonely librarians are responsible.
> Libraries have used 639-2 for decades and there is zero interest within the community to change the policy in this respect. Librarians as a rule do not see 639-2 as a Procrustean bed; the standard has served us well.
Give me a break. I have books in conlangs not represented in 639-2, I have books in dialects not represented in 639-2, and I have books in languages not represented in 639-2. I provide the CIP data which the libraries all take down. But 639-3 would permit me (and other publishers *and* cataloguers) additional flexibility and accuracy. And that accuracy translates not only into comfort for the librarian, but into better opportunity for sales.
It’s bad enough that the Library community still can’t handle Unicode text most of the time. Publishing and cataloguing is not what it was in the 20th century.
> Language code is a part of bibliographic description and indicates the language in which the book has been written.
Not if you exclude languages in 639-3, it doesn’t.
> Note however that for subject description the rules are different. There the libraries can and will apply whatever level of detail they believe is appropriate, using not language codes but subject headings and / or classification codes to indicate the relevant language(s). For instance, the Finnish national bibliography contains 23 references with subject heading "vanha kirjasuomi" (old written Finnish) which does not have a code even in ISO 639-3. Another example is "amerikansuomi" (Finnish spoken by immigrants in the U.S.A.).
The distinction you make between “bibliographic description” and “subject description” is bogus. I publish some books in Volapük, which even has a 639-1 code. Those books are *correctly* described at all levels of the bibliography. I publish a book in Neo, which has a 639-3 code “neu". But it has to be described as “art” because of… what? Inertia? Laziness? Fear?
That’s crap.
>> It is a genuine problem for book publishers who are trying to supply (for instance) the British Library Cataloguing-in-Publication metadata about our books which are in languages supported by 639-3 and not 639-2. I’m supposed to be helping the cataloguers know what language my books are in. I know better than most cataloguers do, right?
>
> Alas, most publishers are not as helpful as you are.
More and more we are all asked to fill out these forms. The problem is that the forms don’t permit ISO 639-3. They’re based on pop-up menus.
(Over on Amazon they don’t even support all of 639-2, And it is IMPOSSIBLE to communicate directly with the technical experts responsible for cataloguing. They refuse to communicate with anyone except as a customer.)
> On the other hand, I do not think that the problem is as severe as you may think, since libraries can use the 639-3 codes you provide in subject description. So if somebody publishes a book in Livonian for linguists and those 25 people who still speak it as the second language, it would be ideal for us if the publisher uses both fiu (to help bibliographic description) and liv (to support subject description).
Your conservatism is very short-sighted. It’s depressing.
Michael Everson * http://www.evertype.com/
|