[xml-h] Linkbases and "Document Enrichment"
ldodds at ingenta.com
Wed Jan 29 16:27:31 GMT 2003
> >p.s. stable linking in general is a bit of a hot topic in the STM area,
> >with features like DOI (http://www.doi.org) and CrossRef
> >being commonly used. These are probably off-topic here though.
> Might not be off-topic. What's STM?
Scientific, Technical, Medical publishing.
In short both libraries and publishers want the material as cross referenced
as possible: the libraries so they can improve their services to users, the
publishers so they can sell more stuff.
DOI (Digital Object Identifier) is basically a URN attached to some piece
of content that can be run through a resolver (CrossRef) that knows the
current location of the material so the punter can be, er, punted off to
the right location. Publishers assign their own DOIs and publish relevant
metadata to the resolution service. Pretty standard stuff, except there's
a tension thats introduced here: the librarians may have an alternate
view of where you should end up.
They'd prefer it that you were directed to where to get the physical copy,
or a preferred document supplier, than necessarily where the publisher
might want to send you. For example the library might have a relationship
with a supplier, or can get material free from certain locations because
they have a subscription. A library may choose to use another resolver thats
maintained locally, e.g. an SFX server (http://www.sfxit.com/) that allows
them to control the endpoints for a link or otherwise offer a menu of destinations.
So a link isn't just a link in this regard, depending on who you are and where
you're coming from you might end up in different locations. It's known as
the "appropriate copy" problem in STM circles.
Another attempt to improve linking in this space involves defining a standard
for encoding citation metadata in a URL (http://www.sfxit.com/openurl/openurl.html)
this makes links slightly more portable by allowing alternative base URLs to
What I find interesting is this mixture of standardised link formats and user
preferences. Especially given the flurry of interest spawned by Jon Udell's
LibraryLookup experiments which a predicated on recognisable metadata
(ISBN) in a URL. While some of the above leaves the user somewhat at the
mercy of a librarian or publisher, I'm interested in letting the user express
their preferences. I rambled a bit about this a while ago
(http://weblogs.userland.com/eclectic/stories/storyReader$372) and today
stumbled across something that offers just that
Probably not really any new tech here, but maybe an interesting perspective of
linking out in the wild.
XML-DEV Blog: http://weblogs.userland.com/eclectic
Personal Blog: http://www.ldodds.com/blog/
"Pluralitas non est ponenda sine necessitate" -- William of Ockham
More information about the xml-hypertext