Thread by @slightlylate

I'm seeing the argument from certain quarters that "working groups are great for getting collaboration and feedback".

This can be true in the breach, but this is not their design function, at least not as practiced in most Standards Development Organisations ("SDOs").

(thread)

To understand why, and to understand why formal chartered working groups _tend_ towards ossification, gatekeeping, and stagnation we need to keep in mind how they come to be and what problems they're formed to solve.

First, recall that standards bodies (SDOs) are first and foremost IP licensing shops. Their job is to collect the rights and IP of member organisations that are emboided in a specified design and re-license that under terms that are "fair" (whatever that means locally)

Big organisations don't feel as though they can just sign away everything carte blanche -- one of their competitors might join a working group that they're a member of and pull in all sorts of stuff they have no interest in giving away into a design -- so WGs are _scoped_

At the W3C, chartered WGs have specified durations (usually ~2-4 years,) and specified deliverables.

Let that sink in.

Before anyone sits down to the very first meeting, the chairs have to agree on _what they're going to deliver_.

You literally have to know roughly all of what you will have wanted to consider in-scope in 2 years when you start out. There's ability to rescope things slightly, but it's painful. Your best opportunity to do that is your next re-charter (hence I argue for short-ish charters).

Now imagine you find out from a partner that they've got a super important problem; something you didn't know about, but would clearly be a fit for the Frobulation WG. What to do? Two options:

1.) Try to shoehorn it into the Frobulation WG
2.) Find some other venue in short run

The first is attractive for a few reasons:

- all the experts are "in the room"; they literally know everything about the space!
- if you can convince the WG that your thing is in scope & everyone miraculously agrees with your proposal, you have a "fast track" to standardisation

But remember, the WG already has chairs, and an agenda, and deliverables, and a fixed date to deliver them by. Your new idea _is randomizing_ from that perspective. And every WG is busy! The work of dotting every "i" and crossing every "t" is vital, laborious, and time-consuming

Trying to do the sort of open-ended investigations of a solutions space that you need to do to iterate towards the right solution is nigh impossible under these conditions. And that's without considering the constituency considerations.

Yes, all of the "right people" (implementers, knowledgable insiders in the space) are "in the room" at the WG, but why? Because their organisations are members. And those folks _tend not to be web developers_....who are the people you're (often) designing a feature for.

Standards groups aren't a fitness function.

Getting every implementer to believe they could implement your proposed design or producing tight spec text *tells you nothing about it's ability to solve a real problem for developers or users*.

This is why you see Chrome engineers preferring to start new explorations in forums like @wicg_. It's an open space that's super easy for web developers to participate in, which also preserves the ability to migrate a design/spec to a WG at the next rechartering opportunity.

I don't set many policies for the Standards Team here, but "incubate first" is the most important of the short list; finding the spaces where you can iterate rapidly, not annoy the "right people" while you sort through rough hewn ideas, and get real developer feedback is *key*

Chartered WGs of course *want* to take web developer feedback into account the way incubations processes can, but they literally don't have the time...and that's by design.

Anyhow, if you're a web developer and want to have a big impact on the future of the platform, comments you leave in repos and discourse threads in @wicg_ shape things at a critical moment: wicg.io

@wicg_ BTW, none of this takes away from the great work that WGs do. They're the big leagues and CGs are the farm teams. We need them both! But we also need to expect reasonable things of each, else we'll always be disappointed.