RE: Extensbility and Versioning Paces

[Paul Hoffman / IMC <phoffman@xxxxxxx>]

At 1:44 PM -0700 9/19/04, Dare Obasanjo wrote:
> Are you putting a stake in the ground and claiming
> that Atom will not have any backwards incompatible
> errata.

That is an option. It may or may not be a good prediction, but it is 
an option for this WG.

> Speaking of which, what is the IETF policy on errata?

There is no policy. If you look at what has been done in the past in 
different WGs, it's all over the map. Change the version number. 
Don't change the version number, just clarify. Add a version number 
to the new version because we forgot it in the old version. 
Mumble/pray a lot.

The advantage that the IETF has over the W3C is that we've been 
around much longer to make many more mistakes. Poor versioning is a 
great example, right up there with poor extensibility.

>  > Whether a refinement, expansion, clarification is a
> >  compatible or
> >  incompatible change is the tricky part. 
>
> It isn't just the tricky part, it is THE MOST
> IMPORTANT PART from the perspective of implementers.

+1

--Paul Hoffman, Director
--Internet Mail Consortium