RE: Stats v2 proposal - for Thursday's meeting

Harald,
  Given what you say about the number of files to introduce a new interface class, is that an argument for not creating a lot of new classes for error events?  For example,  media capture, recording and webRTC could share a common error type definition.  It's certainly cleaner to have separate classes for each spec, but it's not necessary.  

- Jim
P.S.  Trying for cleanliness when you're working in JavaScript is like shaving when you're living in a dumpster.  It does make you look a bit better, but you're still not fit for polite society.  

-----Original Message-----
From: Harald Alvestrand [mailto:harald@alvestrand.no] 
Sent: Tuesday, December 11, 2012 6:27 PM
To: public-webrtc@w3.org
Subject: Re: Stats v2 proposal - for Thursday's meeting

On 12/11/2012 11:01 PM, Martin Thomson wrote:
> On 11 December 2012 13:38, Harald Alvestrand <harald@alvestrand.no> wrote:
>> On 12/11/2012 07:34 PM, Martin Thomson wrote:
>>> On 11 December 2012 07:48, Harald Alvestrand <harald@alvestrand.no> wrote:
>>>> that's impossible to implement due to some WebKit limitations
>>> Would it make a difference if you made the substitution:
>>> s/dictionary/[NoInterfaceObject] interface/g ?
>> Not much of a difference; if I can tell the difference aright, I also 
>> have to do s/^/attribute/, or s/$/()/ (making all the members either 
>> attributes or functions).
> I think that perhaps I wasn't clear enough - if webkit is unable to 
> produce dictionaries (I'm surprised by this, btw), then is it equally 
> unable to produce interfaces?
Each new interface class requires the addition of approx. 3 files to Webkit (IDL, .h and .cc) (at minimum), and causes the generation of half a dozen more, resulting in a significant number of added code bytes.
If the type should be carried all the way through WebKit to the underlying browser, it usually takes about 6 more files to accomplish that.

It's an uncomfortable meeting place between a strongly typed language and a proto-typed one.
>
> The rest of us are not similarly afflicted by this limitation.  I 
> would expect the specification to use dictionaries throughout.
>
Good for you. WebKit's open source, so you're free to contribute a solution.

Received on Tuesday, 11 December 2012 23:39:38 UTC