Re: [apps-discuss] Fwd: New Version Notification for draft-shelby-exi-registration-01.txt

[Carsten Bormann <[e-mail]>]

Actually, the context in which we (at least some of us) are discussing this does not involve HTTP.  So HTTP header fields definitely are not the right way to handle this.  (But even in HTTP, I'd expect coding parameters to be specified in the payload, just like it is done in RFC1950.)

Our use case is about reigning in the exuberant puzzle of choices.  Flexibility is overrated.  Options often cost more money to implement than the optional functionality, and they hurt interoperability.  We want to nail down as much as possible.  I'd expect us to never send an EXI header, but always define that in the media type.  Yes, we might have a couple more media types in the end.  But the point of standardization is to create markets by reducing choice, so that proliferation will be limited.

Grüße, Carsten