Timing Guide for Atom Implementers
Timing Guide for Atom Implementers
Date: Thu, 16 Jun 2005 13:28:43 -0700 From: Paul Hoffman <email@example.com> To: Atom WG <firstname.lastname@example.org> Subject: Timing Guide for Atom implementers
Greetings again. Some people have started asking (mostly off-list) when the Atom format would be "done" so they can implement and, more importantly, deploy. This message gives our best shot at answering the question.
The status of the format document is that the -09 draft is in the IESG evaluation queue. This means that the IESG will consider it soon. In fact, it has been scheduled for their telechat next Thursday. (There are many documents scheduled; see http://www.ietf.org/IESG/agenda.html.) The IESG members are now reading the document in preparation for the telechat, and they may (or may not) be discussing it among themselves.
Next Thursday, the IESG might [a] say "this is fine as it is, no changes needed". It is much more likely that they will [b] say "this will be OK after you make the following changes in the -10 draft" and they will list a handful of changes that range from editorial nits to adding clarifications. There is a real chance they will [c] say "This has some real technical deficiencies, please revise this and show us -10 before we make a final decision."
If [a] or [b]:
Because there are no technical changes requested to the document, implementers can get ready to deploy. The only significant change that will appear in -10 will be that the final Atom 1.0 namespace will be included. Tim is responsible for getting that minted, but we won't do so until we are sure that there is a technically-stable document; we don't expect the namespace creation to take much time. As soon as the namespace is minted, we'll tell the WG. (There will be other changes in -10, such as the last editorial nits, and the addition of the longer acknowledgements list. In fact, even if [a] happens, there will be a -10.)
Also, after -10 is published as an Internet Draft, Scott Hollenbeck will remove the last "hold" on it and it will become an IETF Proposed Standard. Note that it is really a standard at that moment, not when it gets published as an RFC (which will be 3 or 4 months later <sigh>). Implementers can start shipping as soon as it is a standard, and not wait for the RFC.
However, in the case of [c], implementers should put their development on hold. If the IESG asks for a technical change, it will take at least a few weeks for the Atom WG to reply either with the change or with a stack of solid arguments why our original text was, in fact, correct. In the latter case, the IESG will probably take at least a few weeks to agree with us, or for them to say "no, we really meant it". There could be multiple rounds, so the -10 might not be final.
The -09 draft of the format cannot be considered final and deployable. The IESG may ask/demand changes in it before it becomes a standard, and we will not give it a "final" namespace until the IESG says that it won't be changed. At that time, we will quickly get a nice namespace minted, and tell the WG what it is. There is a real chance that the document will be a standard two weeks from now; there is a real chance it will take a few months.
If all this IETF process stuff is confusing, reading the Tao of the IETF (http://www.ietf.org/tao.html) might help. Also, there is a link to the IETF status of the document on the wiki (http://www.intertwingly.net/wiki/pie/FrontPage).
--Paul Hoffman, Director --Internet Mail Consortium