- From: Robin Berjon <robin@robineko.com>
- Date: Mon, 9 Nov 2009 17:34:58 +0100
- To: Frederick Hirsch <frederick.hirsch@nokia.com>
- Cc: "ifette@google.com" <ifette@google.com>, W3C Device APIs and Policy WG <public-device-apis@w3.org>
On Nov 9, 2009, at 17:27 , Frederick Hirsch wrote: > I think we roughly ended up in our F2F discussion with something > like you suggest, two classes of APIs: > > 1. those with implicit user consent, eg. api that opens camera > interface but requires user to press button to take picture, or > messaging interface that requires pressing "send" etc > > 2. those that are wholly programmatic, e.g. message sent without out > user interaction, picture taken without user interaction. > > The second class clearly requires policy controls, the first class > might if decision apart from user's is required or further > restriction is needed. Right, and furthermore there was doubt from some parties as to whether the two APIs could be the same. I think that we should work on the assumption that they can be, and break that only when it turns out not to be possible. In general I think that if we design to the more constraining case (there is no policy control) then it'll work for the more flexible ones (it might look strange at times — for those cases convenience wrappers may help). -- Robin Berjon robineko — hired gun, higher standards http://robineko.com/
Received on Monday, 9 November 2009 16:35:37 UTC