[openbiblio-dev] BibSoup in AnimalGarden

Mon Dec 19 19:58:23 UTC 2011

Peter Murray-Rust <pm286 at cam.ac.uk> wrote:

> > and if not in a general text editor, with what editing tool?
> Is there a tool that does this? 

There are many tools for editing BibTeX or logical equivalent, which is largely adequate as a
BibJSON substrate. e.g. Zotero, Mendeley, BibSonomy, BibDesk, JabRef, ...
Any of these could be used to edit a collection and then publish it to BibSoup. We should
as soon as possible support proper between the native SQLite files used by Zotero and
Mendeley and BibJSON.

> I wanted to edit some myself last night - tried to get a JSON plugin for Eclipse and after 3-4
> hours gave up.

Good JSON editors seem hard to find. I am not sure this is the right way to go either, to use
a general purpose JSON editor. Need some validation to  keep BibJSON on the rails I think.

> I want to suggest that BibSoup and BibServer can:
> * look up books and articles in existing compilations. (We should be able
> to do this with British National Bibliography - I assume most Penguin books
> will be in it?)

Yes! This is an important use case. Users selecting  records they care about from larger collections.
We should similarly plug in to arXiv, DBLP, RePEc and other substantial open sources, and have an architecture
that makes it easy to add more such sources. A typical domain-specific BibServer installation will only
need a few of these. But it is just great for users to be able to easily pop out of their own collections, 
into a larger search space, and then bring some records back for editing and tagging as desired.

> * resolve ISBNs against a Bibliography (again I would use the BNB)

Yes. This is exemplary of general id lookup and response service which Mark and I have discussed and agreed
in principle to build.

> * allow us to annotate the BibSoup with local details (e.g. where the book is)

Good. Need a nice anno tool or popup for this.

> * allow us to map local identifiers (i.e. OWL's id vs Penguin's id)

Supported in principle by the present data structure. Need some matching/deduping tools as well though,
and an interface to confirm/deny matches. I have some starts on this.

> * look up books by publisher, date, author (e.g. all Penguins published in 1991). I had a friend who used to collect first editions of Penguin crime
> (green covers).

This is pretty much there out of the box, provided publisher is specified as a facet. Publisher names are varied though.
Best to work  with a controlled list of all publishers. Other discussion on list about this.

> This is likely to British to start with as we have the BNB and I can use
> simple examples. That's then an incentive for people to get other bibliographies into this.

Definitely. All good suggestions!

--Jim

----------------------------------------------
Jim Pitman
Professor of Statistics and Mathematics
University of California
367 Evans Hall # 3860
Berkeley, CA 94720-3860

ph: 510-642-9970  fax: 510-642-7892
e-mail: pitman at stat.berkeley.edu
URL: http://www.stat.berkeley.edu/users/pitman