On 11/21/11 10:38 AM, Anne van Kesteren wrote:
On Mon, 21 Nov 2011 16:16:22 +0100, Boris Zbarsky <bzbar...@mit.edu> wrote:
As long as all technical discussion ends up in a central place where
everyone can see it at some point, no harm done.
My experience is that once you have side channels for technical
discussion, that doesn't happen anymore. Plenty of stuff gets
discussed on irc and makes it into the spec without any mention on
this mailing list, for example.
The net result is that it becomes easy for small echo-chamber groups
to push through changes to the spec that are bad (whether on purpose
or not) that everyone else is supposed to notice "somehow" and go
about fixing.
You neglect to mention that those changes can also be good
I was specifically addressing the issue of what harm can be done.
Obviously, good changes can come from any source, including benevolent
dictators and random-number generators. ;)
and what the trade off is between the two.
Sure. Again, I was pointing out that there _is_ a tradeoff here, not
just an unmitigated good.
In case a change is made people disagree with it does not take a long
time for it to either be reverted or changed to something that
accommodates even more people. That is my experience thus far anyway. If
your impression is different it would be good to know what we can do to
improve the situation.
My "impression" is that following all changes to the specification via
the revision control system is a pretty large burden, if nothing else
because there is no obvious way to do it linked from anywhere I can
find. Maybe a small set of people "in the know" who got a link from
someone on IRC are following it, but plenty of people who are trying to
implement the specification seem to not be on that select list.
-Boris
