- From: Anne van Kesteren <annevk@opera.com>
- Date: Tue, 28 Feb 2012 18:31:28 +0100
- To: "Jonas Sicking" <jonas@sicking.cc>
- Cc: "Israel Hilerio" <israelh@microsoft.com>, "James Robinson" <jamesr@google.com>, Odin Hørthe Omdal <odinho@opera.com>, "Web Applications Working Group WG" <public-webapps@w3.org>, "Joshua Bell" <jsbell@chromium.org>, "Simon Pieters" <simonp@opera.com>
On Tue, 28 Feb 2012 17:56:05 +0100, Jonas Sicking <jonas@sicking.cc> wrote: > Just to be clear. We've all been here long enough to know that Last > Call is not what matters, what matters is when implementations ship > and when enough content is authored that making incompatible changes > will break too much content. This is an argument you yourself has made > in other specifications and which I agree is reality we have to live > with. I've mostly made that argument when defining legacy features I hope. I do think for new features we should have some explicit call for review when the initial 80% of the work is done. Also, I believe to date no implementation has shipped Indexed DB per specification (that is, without prefix) so that we have this argument at all given that everyone should agree that following Web IDL is important is somewhat surprising to me. > If you are worried IndexedDB hasn't received enough review, I strongly > encourage you to start reviewing. I don't think I'm necessarily the right person, but I've pointed out that I think DOMStringList and numeric constants should go. > I personally only know of one other "bad" thing in the API (the fact > that the second argument to openCursor is a string rather than a > dictionary), but since it's something that can be fixed in the next > version I prefer to fix it there. I'm not sure I'd call that fixing as authors will have to learn the method works in two different ways, but sure. -- Anne van Kesteren http://annevankesteren.nl/
Received on Tuesday, 28 February 2012 17:32:13 UTC