- From: Robin Berjon <robin@berjon.com>
- Date: Mon, 5 Oct 2009 12:53:57 +0200
- To: "Nilsson, Claes1" <Claes1.Nilsson@sonyericsson.com>
- Cc: "'Jere.Kapyaho@nokia.com'" <Jere.Kapyaho@nokia.com>, "brian@westcoastlogic.com" <brian@westcoastlogic.com>, "public-device-apis@w3.org" <public-device-apis@w3.org>
On Oct 5, 2009, at 12:08 , Nilsson, Claes1 wrote: > I agree with Robin and Jere. Don't see the logic of hanging device > APIs off the Navigator object. Actually, to reframe the debate (which is naturally still open), the question isn't so much about logic :) The question is really about how many cows have walked down this particular path [0]. To give a little bit of context for those who weren't on the Web APIs WG in its early days, this is similar to the XMLHttpRequest discussion. XMLHttpRequest is an API that can transmit things that aren't XML, can do so over transports that aren't HTTP, and can do things that aren't actually requests. A case could be made that it's a really poor name, but since it was already in widespread use there was no value in changing it. The case we are concerning ourselves with here is not anywhere nearly as clear cut. Here are the arguments so far in favour of hanging off navigator: - it doesn't pollute the global namespace - geolocation is there - PhoneGap does it - platform (very limited metadata on the underlying OS) - registerProtocolHandler() and registerContentHandler() (which are like AppLauncher but in reverse) (Anne: I'm not counting onLine as that can be controlled at the UA level rather than the complete system, nor language since that's a UA setting even though it may default to what the platform uses for some browsers). That's not a lot of cows but it's some — which is why the decision isn't that obvious. I guess that in general the case can be made that for a web application "navigator" is the layer right below, in the same way that "system" is in a number of language libraries. Keep arguing :) [0] http://www.w3.org/TR/html-design-principles/#pave-the-cowpaths -- Robin Berjon - http://berjon.com/
Received on Monday, 5 October 2009 10:54:30 UTC