[openbiblio-dev] BibJSON vs RDF

Peter Murray-Rust pm286 at cam.ac.uk
Tue Feb 7 23:46:47 UTC 2012

> On 7 Feb 2012, at 22:03, Tom Morris <tfmorris at gmail.com> wrote:
> > On Tue, Feb 7, 2012 at 1:53 PM, Peter Murray-Rust <pm286 at cam.ac.uk>
> wrote:
> >
> >> [bibliographic data from the German National Library] is available in
> RDF as CC0.
> >> Mark will have a look at at - maybe it's easy to convert to BibJSON -
> but
> >> anyway we should work with the suppliers to create BibJSON - this is a
> >> really clear example of why BJ is useful.
> >
> > What is a really clear example?  I missed a step or three in the
> argument here.

It's useful because we can represent the content simply and in a way that
the new tools can display

> >
> > Is BibJSON being proposed as a replacement for RDF?  An alternative to
> > RDF?  What additional value does it provide that justifies this extra
> > expense?

Of course it's not being proposed as a replacement for RDF. I never said
this or remotely implied it.  It provides lots of additional value. That's
*additional*, not alternative.

There's a spectrum of well-designed languages which trade off formal power,
complexity, rigour, ease of use and ease of adoption. RDF is very powerful
and requires a significant toolchain. We used RDF last year in the first
phase of the project - 20 million records from Pubmed, expands to about 4
gigatriples (the NIH pump out a LOT of stuff in each record - BTW they use
XML which is mapped on databases and ASN.1). This was too much for even a
fairly powerful machine running standard RDF tools such as Jena or sesame
(IIRC). So we decided to change to JSON.

There are pluses and minuses to RDF and JSON - I could list them if you
like. JSON fits the role of our project very well:
* good tool support
* widespread adoption
* very browser-friendly
* fairly human-readable and authorable for mid-range geeks
* lots of biblio work already done by Jim Pitman and the BibKN community

There's things JSON is also weak on. Namespaces. Links. But it does at very
good job for us at present.

> >
> > People like Tim Berners-Lee, the W3C, Talis, etc have been pushing RDF
> > to libraries for *years* and are just barely beginning to get
> > traction, so you need to plan on a similarly long and expensive
> > crusade to do the same thing with BibJSON.

By "pushing to" I assume you mean trying to sell them the
concept/technology - not requiring them to ingest it. But no, we don't
intend a long campaign trying to sell BibJSON to libraries. We see the
BibJSON being used outside libraries and for all sorts of purposes. If
necessary we may need to write a RDF2BibJSON converter. Who knows?

> It's one thing to use it
> > as the native format for the BibServer API.  It's quite another to
> > propose it for global adoption in competition with the W3C.
> >
> Don't jump to conclusions! We aren't in competition with anyone - we
intend to work with everybody.  We spent last year working with the
bibliographic community to help get biblio data released in Open form.
That's happening. So there's a great deal of unexplored territory and we
are building the tools that we thing will be valuable.

We are not:
* building a library management system
* building yet-another reference management system
* helping the universities build tools to measure academic performance by

We are making bibliographic tools, data and the fruits of that available to
people who haven't used them before (as wel as those who have) and help
them solve a range of new problems.

> On Tue, Feb 7, 2012 at 10:36 PM, Dan Brickley <danbri at danbri.org> wrote:

> Of course one of the nice things about open data in machine-friendly form
> is that anyone can redistribute it in other formats downstream. I don't see
> any harm in BibJSON enthusiasts doing just that. Or MARC, Excel, ...
> anything else that there's an audience for.
> Exactly

> Dan
> Dan knows I would do everything in XML if the team allowed me to...

Peter Murray-Rust
Reader in Molecular Informatics
Unilever Centre, Dep. Of Chemistry
University of Cambridge
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/openbiblio-dev/attachments/20120207/ee8273bd/attachment.html>

More information about the openbiblio-dev mailing list