- From: Brian Kardell <bkardell@gmail.com>
- Date: Tue, 12 Mar 2013 10:02:56 -0700
- To: Anne van Kesteren <annevk@annevk.nl>
- Cc: Rafael Weinstein <rafaelw@google.com>, whatwg <whatwg@lists.whatwg.org>, "Olli@pettay.fi" <olli@pettay.fi>, Alex Russell <slightlyoff@google.com>, Jonas Sicking <sicking@mozilla.com>, Adam Klein <adamk@google.com>
On Mar 12, 2013 12:06 PM, "Anne van Kesteren" <annevk@annevk.nl> wrote: > > On Tue, Mar 12, 2013 at 3:05 PM, Brian Kardell <bkardell@gmail.com> wrote: > > I was going to mention this the other day - it works inter-operably today, > > so it seems like you probably don't want to break that. Simultaneously it > > does seem to me that the API is more sensible and less confusing - is there > > any reason not change the proposal such that the intent is to to deprecate > > the existing way and consider the new/proposed API as merely superceeding > > the old? Given that one is merely sugar on the other anyway - it should be > > possible to propose the change and augment/prollyfill the mapping I think > > and I see no reason you couldn't quickly roll that out natively given its > > simplicity. > > Yes, that is the basic argument made time and again. It neglects the > costs. It takes away time from people that should probably work on > other stuff, it increases the API surface area and what needs to be > tested, and thereby increases the chance for mismatching functionality > across user agents. Making changes, even seemingly trivial ones, > across multiple independent engines is not something that should be > taken lightly. > > > -- > http://annevankesteren.nl/ Anne, I feel like you've misunderstood my comments/observations. Let me clarify and see: 1) I think this API is more sensible - I had the same problem with mutation observers api. 2) we should not be forever stuck with an unintuitive API 3) we have interop now and a mature spec, that sucks in retrospect that we didn't see this earlier, but it is what it is. 4) this adds nothing in the way of features, it is merely sugar/better API so it should be easily prollyfillable. As such, I am suggesting that we draft a proposal to do so, explain in detail how it would work (I gave a high-level suggestion), provide test cases, etc... That's what prollyfills are all about - a way to evolve outside the browser impl itself based on competition and data which makes that part easier and makes sure we don't take turns that users have problems with ... and ... 5) should we reach that point (it is entirely possible we dont) the actual implementation should be *comparatively *low cost. Does it make sense? Do you feel like I am hand-waving away any of your concerns? I hope not because the idea there is precisely to help address concerns like these (as well as many others). -Brian
Received on Tuesday, 12 March 2013 17:03:25 UTC