Re: Version vs. Namespace

[Paul Hoffman / IMC <phoffman@xxxxxxx>]

At 8:00 PM -0700 7/12/04, Tim Bray wrote:
> If we write Atom 1.0 with a clear guideline that the versioning 
> policy depends on future versions being well-ordered, it will be 
> clear to anyone who produces a later version that isn't that they 
> will be breaking Atom versioning.  Whereas nobody can prevent future 
> stupidity, it would seem a fairly unlikely scenario.

Lots of experience in the IETF reflects this reality. The better the 
version policy is described in the 1.0, the less likely it is that 
someone will do it wrong. The converse has even more examples, 
unfortunately.

> I think that linear version numbers plus MustIgnore hits a huge 
> 80/20 point. -Tim

I would like to see this specifically shown against Mark's scenario 
list from earlier today.

--Paul Hoffman, Director
--Internet Mail Consortium