On May 9, 2011, at 10:29 AM, Jeroen van der Ham wrote:
On 07/05/2011 23:14, Freek Dijkstra wrote:
Compatibility with GLIF and perfSONAR usage
-------------------------------------------
The current syntax is compatible with both current usage, although it is
now specified that recipients of a URN SHOULD NOT interpret the local
part. This is a change from the existing use.
Also, the document specifies that the following two URNs are NOT lexical
equivalent. I have no opinion on this. Should this be equivalent or not?
- urn:ogf:network:example.net:path:2011-0418
- urn:ogf:network:domain=example.net:path:2011-0418
I think the "domain=" part is superfluous in the way that we now
construct identifiers. In order to reduce cruft and burden the urn with
legacy from the start, I say that we make them inequivalent.
I concur with the above. It'd be good, in general, to not have any weird special cases so that comparing URNs is a generic string comparison instead of "does A = B? no. If I change A from domain=, does A match B? ...". There are a few references containing example URNs with the "domain=" aspect. It'd probably just be good to strip them out.