On Tue, Dec 11, 2012 at 9:44 AM, Randell Jesup <rjesup.n...@jesup.org> wrote:
> I'm ok with it,

Great!

> but I think it may be insufficient to avoid problems.

It turned out to be harder to get agreement on the part whose purpose
was to avoid problems: the part about not shipping experimental stuff
on the release channel.

> One could argue if we're going to try to push everyone to move off the
> current prefixing, let's do it once.

If WebGL and WebRTC don’t want to be pushed now, I’d rather stop
vendor prefixing for everything else and let WebGL and WebRTC to come
round at their own pace than see this proposal sink because of WebGL
or WebRTC concerns. (WebGL and WebRTC living in their own worlds isn’t
quite safe, though, because they can leak. For example, we have WebGL
to blame for making the Web platform dependent on little-endian
hardware. Or maybe the real blame should go to TC39 for not giving
WebGL something better in a timely fashion.)

> Are Google, MS and Apple (and Opera) on-board
> with this?

If we require everyone to move in unison, we’ll never make progress.
Someone needs to make the first move and others will follow if the CSS
situation is any indicator. Previous discussions make me expect Opera
and Google to be the most likely vendors to follow us soon if we stop
shipping vendor prefixes.

> Versioning might be best reserved for
> more-complex-than-average APIs and ones where people want to do
> cross-browser experimentation (pre-release).

Considering that versioning has been an anti-pattern on the Web in
general[1], I think we should wait and see how versioning works for
WebRTC before doing it more broadly. (My general expectation is that
it’s bad to introduce APIs in such more-complex-than-average chunks
that the need for versioning arises.)

[1] HTML tried to version itself but browsers ingest all versions
using the same code. Likewise for SVG. On the other hand, XML tried to
version itself but coupled versioning with such a disastrous error
handling policy that their 1.1 launch flopped. JS versioning is
virtually unused on the Web (as opposed to Firefox-internal code). CSS
has been wisely designed as versionless from the processing point of
view.
-- 
Henri Sivonen
hsivo...@iki.fi
http://hsivonen.iki.fi/
_______________________________________________
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Reply via email to