[Standards-JIG] What does 'Draft' status mean?

Peter Saint-Andre stpeter at jabber.org
Wed Jan 11 21:41:07 UTC 2006


Chris Mullins wrote:
> I have a question about the motivation behind changing pubsub.

The motivation behind the two substantive changes was enabling 
implementors to offer simplified pubsub for extended presence etc.

The motivation behind the many readability changes was enabling 
potential implementors to grok the protocol.

> In early to mid 2005, I suggested a number of changes to pubsub, some
> minor, some major. I was slapped down at the time with:
> 
> [PGM - Wed, 23 Feb 2005]
>> These kinds of changes should be done BEFORE the draft status is 
>> reached. Draft status tells implementors that stuff should not
>> really change that much. There are probably more implementations of
>> pubsub than you realize.

I don't recall exactly what Peter Millard had in mind when he referred 
to "these kinds of changes" and I can't speak for him. Naturally, in 
general we'd all prefer not to make major changes to a spec once it has 
advanced to Draft. But we sometimes do that, especially when the domain 
is somewhat unfamiliar (e.g., we've made some non-trivial changes to 
JEP-0124 while it's been in Draft).

> ... and that conversation degenerated into "It's too late to make
> changes to this JEP. Implement it as is." Which I did.
> 
> Now, nearly a year later, (and after implementing the existing JEP
> 60), I see:
> 
>> This is a major revision, just about every line in the spec has
>> been touched, and the changelog is long

See my last message. There are only two major substantive modifications, 
almost everything else was changed to improve readability.

> Now, the changes to pubsub look great. I'm not looking to debate the
> changes. In fact, quite the opposite - I love that the JEP is being
> updated.

Great.

> What I want to know is what changed between then and now? More time
> has passed. There is at least one more working implementation [mine].
> It seems as if the argument that PGM put forth would be even more
> true now than before.

With experience comes wisdom?

> Specifically, what policies should we have with regards to Draft
> status moving forward? Is the current PubSub JEP (which is draft, but
> obviously now in flux) in a state such that I should rebusmit the the
> various suggestions I made last year?

I would not say that JEP-0060 is "obviously now in flux". Based on 
discussions that occurred on this list back in September or so, there 
was consensus to do Simplified Personal Publish-Subscribe (JEP-0163), 
which required the two substantive changes discussed in my last email. 
It's not as if we're modifying the basic pubsub protocol willy-nilly.

As to general policies regarding changes to a spec when it is in Draft 
status, I'd repeat what I said above: in general we prefer not to make 
substantive changes to Draft JEPs, but sometimes that becomes necessary 
if implementation experience exposes good reasons to do so. In the case 
of JEP-0060, the experience was "pubsub is really powerful but if we're 
going to use it for extended presence and such, we need a simplified, 
slimmed-down profile of it". Thus SPPS, and thus the few substantive 
changes we're in the process of making to JEP-0060.

Peter

-- 
Peter Saint-Andre
Jabber Software Foundation
http://www.jabber.org/people/stpeter.shtml
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3641 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://mail.jabber.org/pipermail/standards/attachments/20060111/7d16a0e4/attachment.bin>


More information about the Standards mailing list