- From: Ricardo Varela <phobeo@gmail.com>
- Date: Wed, 28 Apr 2010 15:55:01 +0100
- To: Frederick Hirsch <frederick.hirsch@nokia.com>
- Cc: ext John Kemp <john@jkemp.net>, Robin Berjon <robin@robineko.com>, "public-device-apis@w3.org" <public-device-apis@w3.org>
Just for consideration: wouldn't this make the syntax highlight and contextual code completion/help for functions used by most IDEs a bit of a hard task? Saludos! --- ricardo On Wed, Apr 28, 2010 at 2:55 PM, Frederick Hirsch <frederick.hirsch@nokia.com> wrote: > +1 > > This approach would also enable extensibility of Javascript methods with > additional parameters added later if needed without necessarily breaking > previous usage. Requiring callbacks at the end would not allow this, isn't > that true? > > regards, Frederick > > Frederick Hirsch > Nokia > > > > On Apr 28, 2010, at 9:24 AM, ext John Kemp wrote: > >> On Apr 28, 2010, at 6:00 AM, Robin Berjon wrote: >> >>> Hi all, >>> >>> as you will recall, during the face to face in Prague we discussed the >>> possibility of switching a lot of our API calls (mostly the asynchronous >>> ones) to object literal rather than positional. This was generally perceived >>> favourably but we resolved to first ask Geo why they hadn't opted for the >>> same. Andrei kindly replied[0]. >>> >>> Geo felt that for them, passing options would not be the most common case >>> so that getCurrentPostion(scb) would beat getCurrentPostion({ success: >>> scb}). Upon reflection, I think that our situation is more complex. Some of >>> our APIs are more likely to have options on a very regular, even systematic >>> basis (e.g. Contacts), and it wouldn't hurt to have a little consistency. >>> I'd therefore like to propose that we go for object literal on all of our >>> asynchronous calls. >>> >>> WDYT? >> >> +1 to your proposal. >> >> Using object literal "schema" for all API parameters would be most >> consistent and allows for named parameters to appear in any order rather >> than requiring a developer to get the order correct to call the function >> correctly. The success and error callbacks could then have standard names >> across the API, making it even easier for developers to remember what to do. >> >> Regards, >> >> - johnk >> >>> >>> >>> >>> [0]http://www.w3.org/mid/n2n708552fb1004071110x6d98c6e6t85e83eef0eef52de@mail.gmail.com >>> >>> -- >>> Robin Berjon >>> robineko — hired gun, higher standards >>> http://robineko.com/ >>> >>> >>> >>> >>> >> >> > > > -- Ricardo Varela - http://phobeo.com - http://twitter.com/phobeo "Though this be madness, yet there's method in 't"
Received on Wednesday, 28 April 2010 14:55:35 UTC