Re: WebIDL: how to address the various audiences and constraints?

Hi Art,

On Sep 28, 2009, at 14:40 , Arthur Barstow wrote:
> I am wondering out loud here if it would make sense to split up the  
> Web IDL spec? For example, a functional split e.g. the IDL in one  
> doc, ES 3/5 bindings in a separate doc, Java bindings in a separate  
> doc, etc. Or a core/non-core (e.g. L1/L2) split (I think Maciej used  
> the term "simplification" in one of his emails). Perhaps there is  
> some other split that would be useful.
>
> OTOH, splitting specs can create other problems such as synching the  
> specs, increased overhead for the Editor(s), communication (at least  
> 3 WGs plus TC 39), etc.

As I've stated before, I'm in favour of releasing a v1 that only  
contains the basic stuff that we think is solid. That being said  
reorganisation shouldn't be done at the expense of timeliness but  
rather only if it serves it. For process reasons (and also because  
it's better that way) we need some form of WebIDL that other  
specifications can refer to sooner rather than later.

To put this another way, I think that the only thing that can be  
construed as broken with the way which the current WebIDL is developed  
is the time it takes (and this is certainly not Cameron's fault, if  
you have only one editor who can only dedicate so much of his time one  
can't expect miracles no matter how cool or Australian he may be). So  
while I do like the idea of simplifying and orthogonalising, I'd  
rather it were only done where it helps (or doesn't hurt) timely  
delivery.

-- 
Robin Berjon - http://berjon.com/

Received on Monday, 28 September 2009 12:52:49 UTC