[topicmapmail] Re: Document Object Identifiers/CrossRef

[W. Eliot Kimber]

Daniel Rivers-Moore wrote:

> Anthony Coates wrote:
> > Out of curiosity, does the DOI initiative deliver more than the IETF's
> > DDDS
> > (URN -> URL resolution) would, when applied to "urn:publicId:..." URNs?
> > 
> 
> Speaking from the point of view of very partial knowledge, my
> understanding is that, from the IDF's point of view, a URN would be one
> kind of DOI. ISBN would be another.

In the DOI handbook (available at www.doi.org, naturally) says that they
consider the DOI mechanism to be broader in scope than a URN but
semantically consistent with URNs, that is, a DOI identifies an abstract
thing, not a location. However, they point out that from a practical
standpoint there is no useful difference between these two URIs:

URI with the scheme "doi": doi:10.1234/identifier for my work
URN:                       urn:doi:10.1234/identifier for my work

given that there is no generally-deployed URN resolution mechanism,
meaning that in both cases, you, the client agent, have to have
something special to resolve either form of URI. The DOI provides a
browse plug-in that enables resolution of the "doi" scheme so they don't
see a point in adding the complexity of making the URI into a URN. One
key aspect of DOIs is that a single DOI may resolve to multiple physical
artifacts at any given time--for a given article, for example, there
might be an HTML version, a PDF version, an audio version, and a chunck
of metadata in some format that simply describes the work. The DOI
mapping facility (implemented by CrossRef) allows this multiple mapping
(which, now that I think about it, pretty much translates directly to a
topic representing the work with a number of occurrences--hmmm). 

The DOI mechanism also includes a whole infrastructure for associating
descriptive metadata with DOIs, as well as various policies and
whatnot--you really have to read the DOI handbook to get the full
picture, but they've clearly thought it all out and arrived at
approaches that appear to be fully consistent with the HyTime approach,
for example (although I haven't found any references to HyTime yet, but
I also haven't finished reading the whole handbook, much less the rest
of the materials on DOIs).

The doc is clear though that, were there a generally-available URN
resolution mechanism, it would be useful to use it if it had value for
you as a document publisher.

The DOI mechanism is clearly intended to be bigger than any particular
resolution technology so they've designed it and present it as a
mechanism that can be used with any number of addressing schemes,
whether Web based or not. The real point of DOIs is that they provide
persistent, technology-independent identifiers for intellectual works.

Cheers,

E.
-- 
W. Eliot Kimber, eliot@isogen.com
Consultant, ISOGEN International

1016 La Posada Dr., Suite 240
Austin, TX  78752 Phone: 512.656.4139