This Working Group needs a Secretary

And here I thought the IETF AtomPub effort seemed chaotic.  :-)

At the moment, the way that this working group is operating is that 
people are bringing in new ideas, they are being debated on the spot, 
some make it directly into the document, others are placed on the list, 
and still others go nowhere at all.  We started with a bootstrap which 
did not have a rationale document, and this process hasn't changed.  All 
that has changed is that the W3C has imposed additional chairs which 
increase the workload, and (to me at least) further confuse the flow.

This understandably leads to an overworked editor, appearances of the 
potential for impropriety, and disenfranchised contributors ("hey he got 
his input in, but I did not, what gives?")

The way the IETF AtomPub working group normally operated was that all 
participants could bring forward issues and suggestions (as they do 
here).  The secretary's role was to capture and schedule topics.  The 
chairs had multiple roles: they determined when the secretary scheduled 
the next round, the made determinations of consensus, and they acted as 
the first point of escalation -- more on that at the minute.  The 
editors "normal" role was to make the changes that the chairs directed.

That's how things normally (or nominally) worked.  Now lets look at the 
two significant ways in which things were allowed to deviate from that: 
conflict and opportunistic optimization.

If the secretary wasn't doing his/her job: this could be escalated to 
the chairs.  If the editors didn't do his/her job: this could be 
escalated to the chairs.  If the chairs didn't do their job, escalation 
continued up the <insert-standards-body-here> chain.  But given the 
separations and the checks and balances that this organization of roles 
naturally created, escalation was comparatively rare.

The other way that the process routinely deviated was opportunistic 
optimization.  The secretary can opt to not add to the list things that 
were never seriously discussed -- generally because the person who 
brought up the original issue or suggestion retracted it before it got 
that far and no other proponents emerged.  The chairs can preemptively 
take items from the list (and even ones not on the list) and make 
determinations of consensus at any time.  The editor can also fold in 
changes out of cycle; while nominally this was meant only for 
"editorial" changes, in practice the line is rarely that clear as you 
might think, and in reality editors could make any change that they 
thought would be unlikely to be challenged by the work group.

The sum total of the workload of officers the working group goes up 
slightly with the addition a secretary, but in a way that very much is 
in line with the economics of an all-volunteer effort; in particular the 
workload of the editor and the chairs doesn't increase, and may in fact 
substantially decrease.

The more substantial change is that (1) what has and has not been 
decided, and (2) the process for getting ideas into the queue for a 
consensus call to be made, are both made much clearer.  Additionally, 
people who can't keep up with the firehose of mailing lists such as 
these can instead chose to limit their scanning to scheduled cycles, and 
participate further only when topics of interest to them are being decided.

I could give concrete examples of how this could apply to some of the 
current, at times emotional, discussions, but I hope that the 
application is pretty obvious and my goal with this post to the mailing 
list is to reduce rather than reinforce the level of emotion.

- Sam Ruby

Received on Friday, 24 August 2007 12:18:50 UTC