- From: Jeffrey Mogul <Jeff.Mogul@hp.com>
- Date: Thu, 10 Aug 2006 10:26:42 -0700
- To: Julian Reschke <julian.reschke@gmx.de>
- cc: HTTP Working Group <ietf-http-wg@w3.org>
Julian Reschke writes:
Hmm, I think this would still need to be
transform-info = "identity" | "unspecified" | extension
extension = token
...so it *can* be extended later.
More feedback appreciated... Do people think it's worthwhile for a
server to be able to specify the kind of transform?
Yes.
I think the biggest mistake that protocol designers make is
to force endpoints to make inferences, in the name of "keeping
the protocol simple." In fact, if you oversimplify the protocol
design, the implementations become both complex and unpredictable,
because different implementors tend to make different inferences.
If you believe there is ANY good reason why the client might want to
know what the transform is (and I think you have made that case,
in previous email) then I would strongly support a design that
allows (even "encourages") the server to provide specific information.
I'd also suggest creating the appropriate IANA registry in the
same document, rather than doing this in two I-Ds, since it
makes no sense to have an extension mechanism like this without
some plan about how to use it to maintain interoperability.
You might also want to define some of the initial registrations,
if you know what they are, rather than waiting until people
have adopted conflicting meanings for extension tokens.
-Jeff
Received on Thursday, 10 August 2006 17:27:10 UTC