The IESG ballot process was invented by the IESG to implement its responsibility under RFC 2026 to decide unambiguously about the publication of drafts submitted to the it. The ballot position that causes a holdup is a DISCUSS. A draft can't exit IESG consideration until all DISCUSS ballots have been cleared. Normally, that means that the AD concerned has changed to NO OBJECTION. More rarely, s/he could even upgrade to YES. If a DISCUSS can't be resolved by mutual agreement, there are four possible outcomes.
The draft stays in limbo. This is very undesirable although it sometimes happens. Most likely, this will be when several ADs have deeply held technical concerns but the authors and/or WG strongly disagree or have lost interest. ADs and PROTO shepherds should make every effort to avoid this. We all lose if a draft stays in IESG discussion for a year.
The AD changes from DISCUSS to ABSTAIN (also known as holding one's nose). This means the AD has decided to give up the fight but wants it clearly on record. Since a certain number of YES or NO OBJECTION ballots are needed for a draft to pass the IESG for the standards track or BCP, several ABSTAINs could be enough to kill a document. (Actually, the IESG never kills a document: it sends it back to the working group or authors, saying, "We haven't approved it for the following reasons.")
If the document returns to a second telechat agenda (the sponsoring AD can do this), there is only one DISCUSS, no other ADs have expressed support for that DISCUSS and there are otherwise sufficient YES or NO OBJECTION positions for the draft to pass, then the draft is approved. This is a new procedure the IESG adopted towards the end of 2008.
The IESG chair decides to break the deadlock over a single DISCUSS with an override vote. All ADs then have to (re)read the draft, and vote YES or NO. Two-thirds of YES votes and no more than two NO votes are then needed; otherwise the document is sent back.
When to place a DISCUSS and when not? See the DISCUSS criteria (The tracker captured a little of the timeline of its development). It's very important that ADs stick to these criteria (and non-criteria) to avoid DISCUSSes that don't point to precise, fixable problems.
How to resolve a DISCUSS?
When you ballot DISCUSS, you must also type in your reasons into the tracker, making sure that the reasons are clear and actionable, and that they fall within the DISCUSS criteria.
You should use the "Save and send email" button to save your ballot. This will send email with your DISCUSS reasons to the IESG, the PROTO shepherd, the working group chairs, and the document authors. Consider adding the working group's mailing list to the CC, if the issue is likely to need working group consensus (consult with the responsible AD and/or the chairs if there's a question about whether to do this).
Again, your reasons need to match the DISCUSS criteria. It's also much better if you include a suggested fix. If you already have an idea of what needs to be done to satisfy your concern, it will save a lot of time and effort if you're up front with it. Specifying a fix can also serve to make it clearer exactly what you're concerned about, and cycling on misunderstandings about that is a great time waster.
In some cases, the shepherd and authors will immediately slap their heads, say "Ah zut!" (if they're French, or perhaps "Hostie!" if they're from Québec), and agree to the fix. The responsible AD will then be able to determine whether a new draft is needed, or whether a note to the RFC Editor is enough, and the discussing AD can soon clear the DISCUSS and ballot NO OBJECTION.
In other cases, the discussing AD may need to get into email debate with the shepherd, authors and even the working group in order to get appropriate changes worked out and approved. The responsible AD and the PROTO shepherd need to drive this debate to a conclusion. If the authors or working group seriously disagree with (part of) the DISCUSS reasons, there's a danger of limbo unless the shepherds pay close attention.
The discussing AD is also responsible for keeping the discussion going and moving it to conclusion. It's important to be responsive, and to work in good faith toward agreement. You must strike the right balance between being firm and being obstinate, especially when there isn't support from other ADs for your position. Always remember that you alone are blocking progress on a document that has, by this time, rough consensus of a working group and the community.
All parties involved tend to have many other competing concerns and it's easy for weeks to slip away. Shepherding DISCUSSes to a clear conclusion is a really important part of an AD's and a working group chair's task. And of course it can be an emotional business - it's very important to stay calm and professional.
Once agreement on how to resolve a difficult DISCUSS is reached, it's likely that a new version of the draft will be needed. The shepherds may need to chivy the authors again at this stage - DISCUSS discussions sometimes leave a residue of frustration, and authors may understandably lose their enthusiasm for what they see as rather unwelcome editing work.
The content of this page was last updated on 2012-05-28. It was migrated from the old Trac wiki on 2023-02-17.