----- Original Message ----- From: "Geoff Huston" <[EMAIL PROTECTED]> To: "william(at)elan.net" <[EMAIL PROTECTED]>; "Sandy Murphy" <[EMAIL PROTECTED]> Cc: <[email protected]> Sent: Thursday, March 29, 2007 1:07 PM Subject: Re: [Sidr] action items from IETF68 sidr meeting
> At 04:19 AM 29/03/2007, william(at)elan.net wrote: > > >One other thing to remember is that if URL/URI is registered by > >means of RFC, it needs to have protocol that is also documented > >by RFC which rsync is not > > > I was in a URI WG meeting some IETF's ago and I thought that this was > an historical and not a current constraint. I'm not sure what > transpired with this work, so perhaps the best way forward is to > establish what we can from the documentation. So if what you are > asserting is indeed the case then there should be documentation to > support this. Could you please point me to the IETF document where > the need you refer to above relating to a published RFC for a URI > type is clearly stated? > Geoff The guidelines for URI registration are in RFC4395. s2.3 expects the scheme to have a well-defined mapping onto a namespace or protocol or there to be a good explanation of why not. This is not insuperable. There is an individual submission standards-track I-D for SMB (a protocol which could be loaded with issues of being ill-defined, IPR-laden etc) but does, I think, meet the requirements. Tom Petch > regards, > > Geoff > > > > _______________________________________________ > Sidr mailing list > [email protected] > https://www1.ietf.org/mailman/listinfo/sidr _______________________________________________ Sidr mailing list [email protected] https://www1.ietf.org/mailman/listinfo/sidr
