Re: Shipping Atom products prematurely

[Paul Hoffman / IMC <phoffman@xxxxxxx>]

At 1:15 PM -0400 8/17/04, luke@xxxxxxxxxxxx wrote:
> Aggregator authors will just be forced to implement
> Atom 1.0. Unfortunately, as stated above, Atom 0.3 (and 0.4, 0.5 or any other
> intermediate versions there may be) will still need to be supported as well.

That's not at all clear.

So far, we have only seen the pre-IETF version 0.3 in the wild. I 
have heard of zero reports of implementations of the -00 or -01 
drafts that have been seen in feeds. This is a Very Good Thing. FWIW, 
the 0.3 version doesn't even exist as an IETF Internet Draft.

Any vendor who implements from anything other than the final IETF 
standard will have a host of problems interoperating in the future. 
IETF documents expire after six months, and are usually unavailable 
within nine months. So, if someone releases code based on 
draft-ietf-atompub-format-07.txt, that document will soon become 
unavailable, meaning that someone else cannot determine how to 
interop with it in the future.

Of course, the RFCs we produce will live forever. You can still find 
RFC 1 on IMP software at <ftp://ftp.rfc-editor.org/in-notes/rfc1.txt> 
35 years later.

The Atompub WG *will* have proper and understandable versioning in 
both the format and protocol before we are finished.

Until the final, standardized version, the documents will make it 
clear that implementers should only implement for internal use. If we 
find some implementers creating feeds from pre-standards IETF drafts, 
a bit of quick public education can probably fix the problem.

--Paul Hoffman, Director
--Internet Mail Consortium