Posts from Sunday, September 17th, 2006

W3C Change: Working Groups

Published 12 years, 2 months ago

The second area where I think the W3C could be improved is in how Working Groups are populated and managed.  To a large extent, what I propose is just a re-commitment to existing rules, and isn’t particularly radical.  That doesn’t make them any less important, of course.  Furthermore, this area of discussion doesn’t boil down to one talking point; rather, it boils down to three.

First is this: participants in a Working Group should be productive, or else leave the group, whether voluntarily or otherwise.

This is really already part of the rules, but it’s not very well enforced, in my experience.  I mean that personally, too: between mid-2003 and mid-2004, I contributed almost nothing to the CSS WG.  I didn’t even phone in for teleconferences, let alone contribute to specifications.  Now, as an Invited Expert, the participation rules aren’t quite the same for me as they are for Member representatives, but by any measure, I was deadweight.  I was only on the WG membership list out of inertia.

When the WG’s charter came up for renewal in 2004, the chair asked me if I wanted to stay in the group and start contributing again.  After some reflection, I said no, because I wasn’t going to magically have more time and energy to give to the WG.  To stay would have been dishonest at best, so I left.

Honestly, though, he should have asked me the same question (and been a little more pointed about it) six months previously.  WG chairs should do the same for any member who falls silent.  The actual reasons for the silence don’t matter, because having a WG member step down isn’t a permanent excommunication.  It’s simply an acknowledgment that the person is too busy to be a contributing member, and so leaves the group, whether temporarily or for good.

Ideally, people would voluntarily do this upon recognizing their lack of participation, but not everyone would.  I didn’t, until I was prompted.  WG chairs should prompt when necessary, and even be empowered to place someone on inactive status if they don’t contribute but refuse to step down.  Again, this isn’t a permanent decision, and it isn’t punishment.  It’s just keeping the WG membership list aligned with those who are actually contributing.

This brings me to the second point, related very closely to the first: Working Groups should have a minimum membership requirement.

If a WG doesn’t have enough members to operate, then it needs to be mothballed.  Simple as that.  If you had ten WG members and eight of them went silent, leaving you with only two active members, then it’s time to close up shop for a while.  No WG would ever be permanently shuttered this way:  it would simply be placed on “inactive” status.  Once enough people committed to being contributing WG members, it could be re-activated.  Granted, this would require a re-chartering and all the other things necessary during that process.

I also have to figure that if a WG was in danger of going inactive, some of the group’s members would get involved again.  If not, word would spread and community members would step up to offer their help.  And if none of that happened, then it would be a pretty strong indication that the WG did need to be shut down, for general lack of interest.

Of course, all this requires a WG chair who is willing to hold people’s feet to the fire, to cut inactive members, and to shut down his own WG if there aren’t enough active participants.  But then WG chairs are already required to do a lot of things, and not all of them get done.  Some are trivial; some are not.

The biggest obstacle a WG can face is its own chair, if said chair is abrasive or obstructionist or just plain out of touch.  As things stand, the only way to lodge a complaint against a chair is by working your way up the chain of command at the W3C.  That’s a pretty flat set of rather short chains, though.  In many cases, it doesn’t take a whole lot of steps to reach Sir Tim himself.  And there are even cases where WG chairs are their own bosses, hierarchically speaking, which makes it hard to effectively lodge complaints.

Thus we come to my third suggestion: there needs to be a “vote of no confidence” mechanism for WG chairs.

This is nothing more than a vote by the members of a Working Group:  do we keep our chair, or should he step down?  In this way, the WG itself can decide when it’s time for a leader to go.  I get a little wobbly over the actual vote threshold: should a chair be removed if half the WG votes against him, or two-thirds?  Tough call.  Probably a majority, on the theory that any WG with that many people opposed to the chair is already in deep trouble.

I’m also unable to decide whether I’d have these votes happen automatically, on a set schedule—say, every year right before the March Technical Plenary—or only when a member of the WG calls for one.  Both approaches have pros and cons.  I think my slight preference is for the set schedule, but on the other hand, requiring a member of the WG to call for a “no confidence” vote would be useful, in that the mere call for a vote would serve as its own indication of trouble in a WG, regardless of the vote’s outcome.

So that’s how I’d reform WG membership and leadership:  participants need to be active; WGs need a minimum membership to continue; and WGs should be able to remove their own chairs when necessary.