- From: Graham Klyne <GK@Dial.pipex.com>
- Date: Tue, 29 Dec 1998 19:52:25 +0000
- To: Henrik Frystyk Nielsen <frystyk@w3.org>
- Cc: Ted Hardie <hardie@equinix.com>, masinter@parc.xerox.com, Chris.Newman@innosoft.com, discuss@apps.ietf.org
At 21:38 23/12/98 -0500, Henrik Frystyk Nielsen wrote: >At 12:21 12/22/98 -0800, Ted Hardie wrote: >>4) The content negotiation implied by the document is also not >>workable within the current CONNEG framework, because the set >>intersection model CONNEG uses presumes that the resource is intended >>for a single purpose; it has no provision for a resource that is a >>token, a description, and machine-usable code. In the current >>framework, a device selects among multiple values in a set >>intersection by q-value, not purpose. It can't really select "one for >>this and one for that" in the same operation. > >Unless this is different from HTTP then the q values describe the value on the axis and not the dimension of the axis. q values can be applied to any dimension be it type or some other property. In fact, the negotiation hinted at here only spans the media type. If my understanding of HTTP is correct (and, for the purposes of this distinction, not including TCN), each of the dimensions of variability (indicated by Accept:, Accept-charset:, Accept-language:) is presumed to independently variable. Thus, it makes sense to permit q-values to be determined independently for each such dimension. The 'conneg' framework greatly increases the number of dimensions of variability (through feature registration), and also provides a general framework for constraining variations in one dimension with respect to variations in some other dimension(s). Thus it seems more helpful to apply the q-value unidimensionally to a combination of variable features, rather than independently to each dimension of variability. TCN [RFC2295] adopts a similar approach when describing the source quality of a resource, and avoids allowing a client to indicate quality values associated with feature tags. The 'conneg' framework does not preclude further development of some more fine grained q-value indication, but has not attempted to specify such at this time. #g ------------ Graham Klyne (GK@ACM.ORG)
Received on Tuesday, 29 December 1998 13:50:32 UTC