- From: Kenton Varda <kenton@google.com>
- Date: Thu, 17 Dec 2009 01:54:15 -0800
- To: Ian Hickson <ian@hixie.ch>
- Cc: public-device-apis@w3.org
- Message-ID: <4112ecad0912170154p61b70af1td0f87a0607cf3517@mail.gmail.com>
On Wed, Dec 16, 2009 at 10:10 PM, Ian Hickson <ian@hixie.ch> wrote: > You can do this with HTML5-era technologies in at least two ways: an > <iframe> to a site that provides a postMessage() API, and a WebSocket to > the site that provides a WebSocket protocol. Both can be done with "just" > a URL (much like OpenID). > While those are both possible pieces of the puzzle, I don't think they solve the particular problem I was getting at. The problem is that something needs to keep track of what "virtual devices" the user has encountered on other sites in order to populate the device list -- in other words, the producing and consuming sites need a way to discover each other. I suppose a middleman web site could serve this purpose, but the user would have to place a great deal of trust in this middleman. Having the browser keep track of capabilities seems safer to me. But I suppose I should go with the middleman approach for now, and try to get browser support once I can better demonstrate the usefulness. > > > What would you suggest instead of ".data"? > > > > Maybe ".object"? I don't know what words are already reserved. > > "object" is a reserved word in WebIDL, so it would be better to avoid it. > Also everything's an object, so it's even vaguer than "data"! :-) I'll > ponder the issue, however. Maybe "selection", or some such. > Indeed, my concern was that "data" is too specific in this case. :)
Received on Thursday, 17 December 2009 09:55:12 UTC