Re: Proposition to change the prefixing policy

On Tue, 08 May 2012 09:15:25 +0200, Sylvain Galineau
<sylvaing@microsoft.com> wrote:

> I love Paul but I don't really consider this a 'study', nor is it quite  
> what I
> have in mind.

Sure, that's just the best ready-made thing I found. I agree it would be
interesting to do something deeper.

> What I suggest is that we pick a point on, say, gradient's of
> flexbox's timeline after which implementations would have supported an  
> unprefixed
> version of the features. Then see what subsequent changes we would have  
> handled
> differently, or not at all e.g. I suspect we wouldn't have renamed  
> things as late
> as we did in some cases.

For flexbox, I haven't followed it very closely, so someone else should
probably do this exercise.

For gradients, I suspect that we would have concluded that switching to
the "to" syntax was possible (no ambiguity, both can be supported), but
there is a good chance the balance would have tipped towards not doing it,
which I wouldn't regret, even though I voted for the change at the time.
The change in the orientation and start point of angles probably wouldn't
have happened, and that would have been fine too. I don't think anybody
had strong opinions, and we just changed it because we didn't feel the
weight of the content to be compatible with. Deciding that we would work
in pre-computed rgb would probably have been fine, as it impacted existing
content, but mostly not in a breaking way. Same thing for the change in
what corner to corner gradients mean. So the important things about
gradient would have still happened, and the cosmetic changes that some now
regret would have had a better chance of not happening.

Is this the level of analysis you're considering, or do you want to go
commit by commit / resolution by resolution, and ask several parties if
they would have made a different choice?

Do you propose that we spend some time at the f2f working this out, or
some off-meeting time, or doing it by mail?

    - Florian

Received on Tuesday, 8 May 2012 08:19:03 UTC