Re: A List Apart: Articles: Prefix or Posthack

On 07/09/2010 06:59 AM, Eric A. Meyer wrote:
> At 9:02 PM -0700 7/8/10, L. David Baron wrote:
> If you're telling me that vendors are stuck with keeping
> prefixed-property behaviors the same even when changes need to be made,
> then we need to talk instead about ways to version prefixes:
> '-moz-v1-box-shadow' or some such. Otherwise, we're just begging for
> another case (or many cases) of inconsistent implementations of
> unprefixed properties.

I think this depends on the vendor and on the property.

Microsoft has certain stability guarantees that they have to make.
I'm not sure what they are, but they're tied down to that.

Webkit is more flexible about things, but once they've implemented
something that has widespread use (for some arbitrary judgement of
"widespread"), they have a responsibility not to break those uses.
This has happened with border-image in non-Web applications that are
rendered with Webkit, which is why Webkit has to maintain
-webkit-border-image in parallel with border-image. (Web content,
because it is usually written for multiple browsers, is less likely
to be as dependent on a prefixed property, so is less likely to
trigger this responsibility.)

Mozilla can be more aggressive about having authors update because
Gecko is not a system library. Aside from Add-ons, which have to
be updated for each release anyway, applications that use Gecko
usually ship with their own copy. And web content, as I mentioned,
is usually not as heavily dependent on prefixed properties as
engine-specific apps. Given this, there is less drag on Mozilla to
keep existing behavior. And since Mozilla's philosophy strongly
advocates standardization and browser-agnostic content, even at the
expense of breaking some content, there is enough push to overcome
what drag there is.

So Webkit would (and did) consider a -webkit2-border-image, but
in general tries to avoid this. Microsoft seems likely to version
its prefixed properties if they need to change. And Mozilla probably


Received on Sunday, 11 July 2010 08:24:00 UTC