Re: Change Proposal for ISSUE-126

On 16.11.2010 13:41, Anne van Kesteren wrote:
> On Sun, 14 Nov 2010 12:38:30 +0100, Julian Reschke
> <julian.reschke@gmx.de> wrote:
>> On 14.11.2010 12:35, Anne van Kesteren wrote:
>>> On Sun, 14 Nov 2010 12:17:56 +0100, Julian Reschke
>>> <julian.reschke@gmx.de> wrote:
>>>> On 13.11.2010 22:49, Maciej Stachowiak wrote:
>>>>> I don't have strong data on the compatibility impact of this specific
>>>>> change, but when we diverge from both Firefox and IE, it is rarely on
>>>>> purpose, and matching them has almost always fixed bugs, even if we
>>>>> didn't know it at the time.
>>>>
>>>> I'm confident that it really doesn't matter in practice, in which case
>>>> we should default on being consistent with the base specs.
>>>
>>> Because you are confident? Specifications are at the bottom of the
>>> priority of constituencies. Better safe than sorry in my opinion.
>>
>> But how do you know it's safer when UAs currently do not agree on
>> processing, and seem to get away with it?
>
> As Maciej indicated, they are starting to converge.

Well, it's not clear that they *need* to converge (on a spec violation), 
and also not whether what they converge on is the right thing to do.

>> Sometimes it *really* doesn't matter, in which case considerations
>> like consistency and re-usability of code should be taken into account.
>
> It is funny that you say that while also advocating having separate URL
> processors throughout the code.

Do I? If you're referring to the URI-in-HTML permathread: my point of 
view is that we're confusing an pre-processing step with the actual parsing.

Best regards, Julian

Received on Tuesday, 16 November 2010 12:54:08 UTC