ietf-822
[Top] [All Lists]

Re: Time to act: draft-klyne-msghdr-registry

2002-02-12 09:51:25


Dave Crocker <dcrocker(_at_)brandenburg(_dot_)com>
> Writing a spec that has one's name on it very much IS a barrier.

Are you saying that you want anonymous specs, or that you want new header
field names without associated specs?

Speaking as someone who's probably unqualified to talk here (as I've never tried to write an RFC), the mere thought of writing an RFC fills me with terror ;-)

Maybe that's what he's talking about. If I had the idea for a brand new wizzo header field (which I don't..) then I'd be much happier to write a 'normal' spec, at least at the initial stage. You know, maybe half a page or so of 'normal' English to say what the header does, what the syntax is etc. This would be a lot less offputting that writing an 'RFC' style spec where you seem to need 30 pages of "pointless" stuff around the actual "meat" of the document just to put it in the "proper" format, with the fear that if you get it wrong you'll get the spec laughed out of court even though it might be a good idea, just not formatted properly.

Then, if the consensus is that it's a good idea, maybe either someone else, with more experience, will take on the task of writing the RFC style spec, or at least help the original person through the initial problems of writing the spec in the proper format.

To me, a suitable spec would be something like:

List-Unsubscribe
-----------------
Syntax:
 List-Unsubscribe: "<"URL">"[,"<"URL">"]...
where URL is the URL to go to to unsubscribe from the mailing list.

Example
List-Unsubscribe: <mailto:ietf-822-request(_at_)imc(_dot_)org?body=unsubscribe>


That's it.
Now, maybe this isn't precise enough to please some people, but it'd certainly do for a start, and be a lot less frightening than the current requirements (see RFC2919 for the similar List-ID spec)

Anyway, that's my 2p.


> repeat 10 times:  this is a registry, not a standards process.

Yes? The latest draft I see requires that each new registration contains a
"defining document". Old names seem to be grandfathered in by a host of
references, although that isn't said explicitly.

--Arnt