[open-bibliography] Extract Translations from WorldCat?

[John Salter]

Hi Tom,
The xISBN API already returns multiple language variations of an item (and editions) e.g.
http://xisbn.worldcat.org/webservices/xid/isbn/0596002815?method=getEditions&fl=*&format=json (also:
http://www.oclc.org/developer/documentation/xisbn/response-details)

This would cover *recent* items (e.g. items with an ISBN), but not older texts - and I'm not sure how you tell which language is the original?

Does that help at all with what you're trying to do?
Cheers,
John

________________________________
From: open-bibliography-bounces at lists.okfn.org [mailto:open-bibliography-bounces at lists.okfn.org] On Behalf Of Cheesman T.
Sent: 22 October 2012 14:41
To: open-bibliography at lists.okfn.org
Subject: [open-bibliography] Extract Translations from WorldCat?

Hello, dear Open Bibliography group,


[I sent the following to the ScraperWiki list yesterday and David Jones has suggested I try you - saying: "I think some people there have done some WorldCat hacking". But come to think of it, there might be better sources than WorldCat for the data I'm interested in?]

Here's a major challenge - possibly new (?): scrape out of WorldCat all the items which are translations, and link them with all the items which are translated.

Doing this would (1) help anyone find any translations of a work or find the original of any translation (= a tool which performs as an interface to WorldCat data) and (2) build a dynamic dataset representing global cultural flows (= a stats tool).

Data about WorldCat items includes OCLC number, title, date, publisher, language, edition/format, personal names under 'author' (sometimes also 'responsibility'), sometimes 'other titles', 'notes' etc. - the fields vary depending on the library catalogue or other source they came from.

In the case of translations, the fact that a work is a translation is often only shown in the title field:  '... translated from ... by ...' can appear.  So we might have to start by collating all items with such phrases (in all languages) in the title field, and extracting names. Translator names will be key links.

An 'other title', esp if it is in a different language (but that won't be marked), is probably an original title (but might be the title of a translation...).

WorldCat entries are cross-referenced to ''all editions and formats' - and that includes multiple copies, differing editions, and translations, but translations are only linked IF they have the same title as the original (or partly the same). So the English translation of Kristeva's "Hannah Arendt" biography is linked to the Japanese and Polish translations, but not to the French original.

Index Translationum, started in 1932 by the League of Nations, now run by UNESCO, keeps a record of all the world's literary translations. Its records could be used to cross-check links, perhaps?

Just an idea..!

Tom


Dr Tom Cheesman
Reader in German
College of Arts and Humanities
Swansea University
SA2 8PP
T: +44 (0)1792 205678 x 4030
M: (+)(0)7736408064
Institution: www.swansea.ac.uk/staff/academic/ArtsHumanities/cheesmant/
Version Variation: www.delightedbeauty.org and www.delightedbeauty.org/vvv
Asylum: www.swanseabassgroup.org
Publishing: www.lulu.com/hafan


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/open-bibliography/attachments/20121022/3467d15c/attachment-0001.html>