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

Re: DOI and the non-IETF tree

From: <hardie@qualcomm.com>
Date: Fri, 12 Sep 2003 11:44:38 -0700
Message-Id: <p06002000bb87c1259249@[]>
To: "Ray Denenberg, Library of Congress" <rden@loc.gov>, "John Cowan" <jcowan@reutershealth.com>, "Larry Lannom" <llannom@cnri.reston.va.us>
Cc: <uri@w3.org>

At 9:56 AM -0400 09/12/2003, Ray Denenberg, Library of Congress wrote:
>From: "John Cowan" <jcowan@reutershealth.com>
>>  Larry Lannom scripsit:
>>  > 1. The distinction between IETF and non-IETF source and or continued
>>  > control of a scheme seems reasonable, but what is mking of
>>  > building it into the label? That is, what do I know about, say,
>>  > org-doi:10.123/456 that I don't know from doi:10.123/456?
>>  Because the IETF may wish to define a scheme called doi in the future.
>Wow! I had the same question in mind, as Larry, and was anxiously awaiting
>an informed answer, and I suppose this is an informed answer.  I must say,
>in my opinion, it's  quite unreasonable.....

The criteria the IETF uses for defining new schemes in the IETF tree
are set out in RFC 2717.  From the Introduction:

    A registration process is needed to ensure that the names of all such
    new schemes are guaranteed not to collide.  Further, the registration
    process ensures that URL schemes intended for wide spread, public use
    are developed in an orderly, well-specified, and public manner.

That is, avoidance of collision is a major motivating factor for
having a registration scheme.   I would personally find it unlikely
that the IETF would approve a scheme that seemed likely to
conflict.  It is likely that the IETF will try to get folks to register
their schemes as early as possible under some tree, so that the
discovery of potential conflict can occur before the deployment.
That was why I asked Larry to take up the non-IETF tree document
and move it forward.

I think some of the comments made to date have helped scope that
work, and I encourage them to continue, but the goal remains
avoiding collisions and identifying the token holder for change
processing where that is not the IETF.

>....for a number of reasons:
>1. nobody's going to pay attention to the prefix anyway so what this means
>essentially is that DOI would go to all the trouble to register  'doi'
>(whether it's called 'doi' or 'vnd-doi') and have to worry that someday IETF
>is going to register 'doi'  rendering DOI's effort useless.

 From the IETF perspective, these identifiers are for protocol processing,
and we can get the processors to pay attention to the strings.  If
the strings org-idf-doi: and org-isakmpc-doi: are both called DOI by
their respective communities of use, the protocol processor can
tell the difference even if a human wandering into the room has to
ask whether we discussing domains of interpretation for key management
or documents and their metadata.

					Ted Hardie
Received on Friday, 12 September 2003 14:45:17 UTC

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