Re: [ACL] Re: Last minute ACL stuff (was DAV:unauthenticated usage)

At 7:28 PM +0200 04/19/2004, Julian Reschke wrote:
>Yes, I indeed was talking about AUTH48, and I'm aware of the implications.
>
>In this case a property that everybody thought was optionally 
>changeable indeed wasn't. Removing the requirement that it's indeed 
>non-writeable (= protected) doesn't change anything for clients and 
>servers that don't want to support authoring; but it makes it legal 
>for those who wish to do so. As a matter of fact, people that need 
>the ability to change that property will likely support that in 
>their servers no matter what the specs says about it, and thus this 
>would end up on a future issue/errata list anyway. Therefore I think 
>this is a good change that won't have any negative side effects.
>
>On the other hand, I've been announcing that change because it *is* 
>a change, and because we need feedback and full consensus on it. If 
>you or the RFC-Editor feel this one is problematic, I'll be happy to 
>roll back that change (= remove it from the issues list sent back to 
>the RFC Editor).

It is not appropriate to make substantive changes to the document at this stage
of processing.  If this is a showstopper issue, you can ask the RFC Editor
to stop processing the document, and get the WG chairs to call for consensus
on the changes.  The draft will then have to go back through at least
IESG processing and possibly IETF last call.

Continuing to tweak the documents after they have completed processing
is hindering this group's ability to get a stable specification.  It is normal
to find things as people implement and deploy, but if you never issue
a final spec the number of people actually working with the documents
will remain low.

		regards,
			Ted Hardie

Received on Monday, 19 April 2004 13:53:36 UTC