Re: associating constant feature sets wi
[email protected] (Reynolds Franklin NRC/Boston)
| Newsgroups | gmane.ietf.medfree |
|---|---|
| Organization | Nokia Telecommunications |
| Message-ID | <[email protected]> |
I agree with Koen, that a generic or transport independent negotiation protocol may have only limited value. On the other hand, I would like to see a capability negotiation protocol for HTTP. I am not sure whether the work should be done in this group or the HTTP group. I thought that "delegation" referred to a mechanism for asking one or more third parties to provide partial or complete profile information. Partial profile information from multiple sources could be aggragated in some way to provide a complete profile. Such a third party might be a cache or proxy or some other server not normally part of the communication channel. In the CC/PP paper there was a notion of "indirect references" which were URLs that named profiles with default values that could be superseded by a user agent. I agree with Graham that we can separate most of the issues of naming and representation of feature sets from "transfer". I am very interested in an HTTP specific version of the negotiation protocol but I don't have a strong preference about where the work takes place. Franklin Reynolds Nokia Research Center