Nat, Understood. The reasoning (I assume, since I didn't write the spec) is that KVF is handed back in the body of the reply and/or processed locally (e.g., signature generation), whereas openid.* parameters are handed back via a URL. When sending parameters along with other parameters in the URL, having something like an "openid" namespace makes sense so as to prevent name collisions.
I assume the reason is nothing more than that distinction -- use the namespace form when collisions are possible. Paul > -----Original Message----- > From: Nat Sakimura [mailto:[email protected]] > Sent: Sunday, February 07, 2010 10:20 PM > To: Paul E. Jones > Cc: [email protected] > Subject: Re: "openid." name space of KeyValue Form > > Hi Paul, > > All the libraries that I know do not put openid. in KVF. > The current spec is written like that and it is the correct behavior. > > Until Hideki's message, it never occurred to me. > > I suppose Hideki's is asking the reasoning that we made this asymmetry > in the URL encoding and KVF. > > =nat > > (2010/02/07 21:22), Paul E. Jones wrote: > > For use in Key-Value Form, I didn't see it as necessary when I > implemented > > the spec. It seemed logical not the be there. > > > > The only reason why one might want to use this is to include some > kind of > > non-standard information. Is that something folks would want to > encourage? > > Anyway, changing the spec to have "openid." there now would break > things, so > > I would not recommend it unless there was a really good reason. > > > > Paul > > > > > >> -----Original Message----- > >> From: [email protected] [mailto:openid-specs- > >> [email protected]] On Behalf Of Nat Sakimura > >> Sent: Monday, February 01, 2010 12:14 AM > >> To: [email protected] > >> Subject: Re: "openid." name space of KeyValue Form > >> > >> Hmmm > >> > >> That's a good question. The reason we put openid.* in the request > and > >> response is that there may be other applications sharing the same > >> request/response. If so, it would be more consistent if we put > openid.* > >> prefix to the keys of the direct response as well... > >> > >> Is it just an oversight, or did it have a good reason for it? > >> > >> =nat > >> > >> (2010/02/01 13:49), nara hideki wrote: > >> > >>> Hello all, > >>> > >>> I'm thinking of the good reason why "openid." name space to keys of > >>> Key-Value Form Encoding used for direct responses is dropped. > >>> I think that we MAY use "openid." name space. > >>> > >>> I'm very happy if someone give me a good cue to understand the > >>> > >> reason. > >> > >>> Thanks in advance. > >>> ---- > >>> hdknr > >>> _______________________________________________ > >>> specs mailing list > >>> [email protected] > >>> http://lists.openid.net/mailman/listinfo/openid-specs > >>> > >>> > >> > >> -- > >> Nat Sakimura ([email protected]) > >> Nomura Research Institute, Ltd. > >> Tel:+81-3-6274-1412 Fax:+81-3-6274-1547 > >> > >> _______________________________________________ > >> specs mailing list > >> [email protected] > >> http://lists.openid.net/mailman/listinfo/openid-specs > >> > >> > > > > > > _______________________________________________ specs mailing list [email protected] http://lists.openid.net/mailman/listinfo/openid-specs
