W3C home > Mailing lists > Public > public-device-apis@w3.org > August 2009

Re: ISSUE-4 (api-versioning): API Versioning [APIs - General]

From: Marcos Caceres <marcosc@opera.com>
Date: Wed, 26 Aug 2009 13:09:09 +0200
Message-ID: <b21a10670908260409i28ede6eataeb91c8a9ed7b4d3@mail.gmail.com>
To: Anne van Kesteren <annevk@opera.com>
Cc: Robin Berjon <robin@robineko.com>, "<richard.tibbett@orange-ftgroup.com>" <richard.tibbett@orange-ftgroup.com>, public-device-apis@w3.org
FWIW...

On Wed, Aug 26, 2009 at 12:46 PM, Anne van Kesteren<annevk@opera.com> wrote:
> On Tue, 25 Aug 2009 16:54:11 +0200, Robin Berjon <robin@robineko.com> wrote:
>>
>> I'm in full agreement. I'd simply add that "approved" should normally mean
>> that it's reached Recommendation status, but can at times be extended to
>> apply to something that is really widely deployed and used (e.g. this has
>> been a guiding principle in specifying the XMLHttpRequest API, even though
>> it meant that it's not all that elegant everywhere).
>
> With XMLHttpRequest we are making some changes though that some could
> consider "breaking". E.g. with level 2 we're overloading send() more, and
> open() no longer throws when passed a non same-origin URL.

Agreed, so long as the the original method still works and does not
break old content.

> I think such changes should be acceptable. Making changes because
> implementations all differ from the specification and are unlikely to change
> should also be ok I think, if not encouraged.

Also agreed.

-- 
Marcos Caceres
http://datadriven.com.au
Received on Wednesday, 26 August 2009 11:10:16 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 14:53:38 UTC