brendan at mozilla.org
Thu Jan 22 16:11:07 PST 2015
Brendan Eich wrote:
> The reason to label editions or releases is not to give marketeers
> some brand suffix with which to tout or hype. It's to organize a
> series of reasonably debugged specs that implementors have vetted and
> (partly or mostly) implemented.
> I agree it would be best if (partly or mostly) were "fully", but
> that's not practical with big "catch-up" specs. With "rapid-er
> release" annual editions, it should be a goal, IMNSHO. That's the
> promised land we seek: implementor- and developer-tested draft matter
> that "sticks" and *then* gets the de-jure stamp of approval.
The WHATWG "living spec" alternative eschews any series of spec
snapshots, favoring just a bleeding edge that implementors constantly chase.
This ideal has real-world issues! Perhaps Boris Zbarsky or someone else
will comment on them. I'm out of time and not motivated, since for JS,
we will promulgate an evolving series of spec editions, from ES6 =
ES2015 onward at an annual cadence.
Part of the benefit of the cadence, which resonates with faster software
rapid-release schedules such as Chrome's and then Firefox's: you avoid
schedule chicken by waving off anything unready till the next "train".
There's always another one coming, and no way to delay, so it doesn't
pay to pretend to be more "done" than you really are.
The rapid-release approach still requires skill and art to get right and
avoid missing a train (always a set-back and embarrassment, as in real
More information about the es-discuss