Lists Home |
Date Index |
John Cowan wrote:
> IMHO the urn:newsml: scheme is a huge win in this respect. The
> details are at http://www.ietf.org/rfc/rfc3085.txt , but the
> general scheme is this:
Thanks for reminding me of this. I've been telling myself for
days that I should propose that the Atom spec include a reference to
this technique of URN generation in order to provide guidance to folk
in generating the unique IDs that Atom *requires* for each entry.
The method used in newsml is easy to use, easy to explain,
cheap to compute, and even deals with such bizarre problems as
ensuring that these URNS are unique 2,000 years from now when someone
else might own your domain name, etc....