W3C home > Mailing lists > Public > uri@w3.org > September 2003

Re: DOI and the non-IETF tree

From: Larry Lannom <llannom@cnri.reston.va.us>
Date: Fri, 12 Sep 2003 09:55:47 -0400
Cc: uri@w3.org
To: "Daniel R. Tobias" <dan@tobias.name>(by way of Martin Duerst <duerst@w3.org>)
Message-Id: <CF901694-E528-11D7-AF63-000393011BFE@cnri.reston.va.us>

This has been discussed inside the IDF and there are strong opinions 
pro and con.

One issue is the current mechanics of URN. As I understand it, 
resolving urn:doi:10.123/456 would get you one or more transformation 
rules, in the form of regular expressions, that could be applied to the 
string. Will now have to give a little context on DOIs -- All DOIs are 
currently registered in the Handle System (informational RFCs about to 
come out, I believe, but see www.handle.net for details). Resolving a 
handle gets you sets of typed values (skipping lots of detail that I'm 
not sure is appropriate to the list). The great majority of registered 
DOIs have for now only a single useful value, outside of admin stuff, 
and that value is a default URL. So they function now as single level 
of indirection -- if you want to know about this DOI, go to this URL. 
That same vast majority of DOIs are now referenced through a 
http-to-handle proxy or gateway service that returns http re-directs, 
e.g., http://dx.doi.org/10.1038/35057062 will return a re-direct to the 
web page for the journal article identified by that DOI.

So this would fit the urn pattern I think, where urn:doi: would turn 
into the address of the proxy server and could be changed later or we 
could even have another proxy that returned a page full of alternate 
links, and so on.

What is evolving and now being prototyped, however, is a more elaborate 
data model which will provide multiple pieces of current state data re. 
the given DOI, e.g., execute this program at this location to get this 
kind of information or result, go here for core metadata, contact this 
organization for more information, license this thing in this way, 
etc., which we expect to be used by various applications. The urn 
approach doesn't seem to fit this.

The second, more general issue, is whether registering an identifier 
such as DOI as a URN means that it IS a URN or simply means that URN 
can be used as one of several interfaces to the underlying identifier 
system. Practically speaking, would registering DOI as a URN preclude 
registration as a URI or some other type of interface?


On Friday, September 12, 2003, at 07:27  AM, Daniel R. Tobias (by way 
of Martin Duerst <duerst@w3.org>) wrote:
> On 11 Sep 2003 at 20:54, Al Gilman wrote:
> > DOI sounds like a second run at defining URNs that seems to have
> > acquired a following and demonstrated interoperable practice.
> But shouldn't it be "urn:doi:" instead of just "doi:"?  If it's URN-
> like, why not make it an actual URN?
> --
> == Dan ==
> Dan's Mail Format Site: http://mailformat.dan.info/
> Dan's Web Tips: http://webtips.dan.info/
> Dan's Domain Site: http://domains.dan.info/

Larry Lannom
Corporation for National Research Initiatives (CNRI)
Suite 100, 1895 Preston White Dr, Reston, VA 20191

email:  llannom@cnri.reston.va.us
tel:  703 620 8990
Received on Friday, 12 September 2003 09:55:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:06 UTC