Re: A List Apart: Articles: Prefix or Posthack

On Jul 9, 2010, at 10:28 AM, Boris Zbarsky <bzbarsky@MIT.EDU> wrote:

> On 7/9/10 10:17 AM, Brad Kemper wrote:
>> I didn't say you could never change the way the experimental property works. But by the time it reaches CR, it's got a lot of momentum behind the prefixed versions in the wild
> 
> How did we get to CR without prefixed versions that match what the CR draft says and thus testing of said prefixed versions in the wild?
> 
>> Thus, Apple supports both "-webkit-border-image' and 'border-image', with significant differences
> 
> Right, and depended on other UAs to do the in-the-wild testing of the new border-image syntax, yes?  But you can't have all UAs thus depending on other UAs, because then none of them will actually be testing the updated syntax.

I expect a property to be much more subject to change in the early days of experimentation than it is nearly a decade later, as with some of these.  For later changes that are significant enough to break designs, there are nightlies and betas that will often be good enough because most of the changes at that point will be more limited usually. In cases where more massive widespread testing is still needed using release versions of browsers, I have stated before that some sort of version number in the prefix would help, such as "-webkit-2-border-image'. But with border-image, the WG voted and said the spec was mature enough to go into CR and thus authors could begin using without the prefix.

Received on Friday, 9 July 2010 18:05:09 UTC