RE: Specifying window.performance.now()

> I don't actually see a need for this entire paragraph, since it's all covered in IDL already.  In 
> fact, I'd prefer not having prose that duplicates the IDL...

I'm okay with removing it.

> Are there security concerns about subframes being able to extract "now" 
> information relative to the navigation start of the root document?

Yes, I believe Sigbjorn and Tony mentioned something similar on the list. We are planning on using navigationStart as the time base origin; navigationStart is unique for every document, even same origin documents.

> I'm not sure what the "MUST never be zero" requirement means in practice.  If the return 
> value is only accurate to 0.1ms, then it seems quite possible to get it twice in a row and end 
>up with the same value on modern hardware, but the text seems to prohibit that behavior.  
> Am I missing something?

Considering monotonically increasing  is defined as for all x and y such that x ˜ y, one has f(x) ˜ f(y), we shouldn't restrict subsequent calls from having zero delta. I will remove that requirement.

Thanks,
Jatinder

Received on Thursday, 1 March 2012 16:40:57 UTC