+1 to unifying, and to be honest I like Chris' suggestion most. I don't particularly care whether we go with an experimental or a registered one, but I think at style of URN will make it very clear that the namespaces do not resolve to anything.
G On 12-09-25 11:14 AM, Christopher Brooks wrote: > On Tue, 25 Sep 2012 19:07:17 +0200 > Tobias Wunden <[email protected]> wrote: > >> Looking through the code base one can see that we now have XML >> namespaces where needed, which is good. However, it turns out that we >> have not been too consistent with regards to how we create these >> namespaces. Some of them are prefixed with http://, others aren't. >> >> Therefore, the first part of the proposal is to unify the two >> approaches and I hope to get a +1 from everyone on this. > > +1 > >> The second part is to agree on the namespace naming scheme we use. On >> the table is the choice between going with a prefix of http:// vs. >> going without it. Technically speaking, the namespace should be a >> URI, which both URN (name) and URL (location) comply with and there >> is no requirement on the prefix, so >> >> engage.opencastproject.org >> >> is as good as >> >> http://engage.opencastproject.org >> >> with the slight difference that using the http:// somehow implies >> being able to visit that location and get something meaningful, which >> is not the case with the Matterhorn namespaces. I am looking for >> reasons to go with one or the other, but to make voting easier, I >> will propose to go *without* http://. > > I support without, but would like it to not just be unique but a URN > instead. The syntax for a urn is: > > urn:<NID>:<NSS> > > The namespace id (NID) must be registered with IANA, but I believe this > is free: > > http://tools.ietf.org/html/rfc3406#appendix-A > > So a completed unr would look like: > > urn:opencast:engage > > We can use, without registration, the experimental URNs. No guarantee > of being collision free, but it's unlikely to cause a collision. That > would look like: > > urn:X-opencast:engage > > I am fine with either of these. > > If we are using http then it should resolve to something, at least a > wiki page. > > Chris >
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Matterhorn mailing list [email protected] http://lists.opencastproject.org/mailman/listinfo/matterhorn To unsubscribe please email [email protected] _______________________________________________
