RE: [websockets] Reintroducing onerror

On Tue, 7 Jun 2011, Adrian Bateman wrote:
> On Tuesday, June 07, 2011 10:36 AM, Ian Hickson wrote:
> > On Tue, 7 Jun 2011, Adrian Bateman wrote:
> > > We have removed onerror from our implementation since the previous 
> > > change and it's frustrating trying track against the spec with 
> > > unexpected updates.
> >
> > That's the cost of being on the bleeding edge. It's very likely that 
> > it will change again. Until we have implementation and authoring 
> > experience, we will almost certainly continue to update the 
> > specification to address new problems and use cases. If you want to 
> > implement things that don't change, then wait til someone else has 
> > implemented it and shipped it. Being on the bleeding edge means you 
> > are providing the feedback that affects the specification's 
> > development.
> 
> I'm just asking for a little transparency here: as far as I can see 
> there were no bugs filed or public e-mail discussion on this nor was 
> there any explanation in the check-in to change the spec. We understand 
> that things change but we're hoping that we're getting close to Last 
> Call and we'd like an opportunity to participate in the discussion 
> before things show up unexpectedly at this point.

That's fair enough. The most recent batch of changes (which are now 
complete, hopefully) were all done as a result of direct conversations 
between Ian and myself, in the context of updating the specs to be in sync 
again. There wasn't really much discussion, it was more just me telling 
him what was wrong with the protocol spec and him telling me what was 
wrong with the API spec, and us then fixing those problems. :-)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Tuesday, 7 June 2011 19:15:52 UTC