[rfc-i] RFC citations committee I-D issued
ted.ietf at gmail.com
Fri Feb 11 10:11:04 PST 2011
Several folks asked what the utility of URNs are here, since they are
generally not resolvable in browsers.
The general utility is that they provide stable identifiers with a
known assignment policy. In the case of
the example I gave before, it tells you that the identifier relates to
an IETF namespace called "id". This
enables you to distinguish between identifiers minted by different
bodies, since they have different
namespaces. So if some other body starts issuing documents whose
names are of the form draft-name-topic-version, you know they are
the same as
How much do you have know about the assignment policies of the two
domains to be sure?
URN namespaces are scoped and provide a way to be sure of what
assignment policy is
in place. Those principles are fairly well adopted by the library
community and you'll note
here: http://www.doi.org/factsheets/DOIIdentifierSpecs.html that DOIs
mimic URN principles
completely despite having chosen not to register as a NID.
For citation purposes, I think you have three choices: reference a
draft by assigned
string (and potentially version), knowing that there is some risk of
conflict with documents
issued outside the IETF; reference a draft by URL, knowing that it may
to bit rot; reference it by stable identifier, knowing that it will
force a search for the identifier
rather than simple resolution. To me, one and three currently
require a search and the trade-off
between them is made depending on the likelihood that someone will
issue draft strings
similar to the strings we use. For IETF, IAB, and IRTF streams, that
risk seems small.
So the tradeoff is really for individual documents. Since two
organizations I have been
associated in the past used draft-name-topic-version for internal
I personally think the 3rd way is better. YMMV.
More information about the rfc-interest