- From: Charles Pritchard <chuck@jumis.com>
- Date: Mon, 04 Apr 2011 00:06:23 -0700
- To: Arthur Barstow <art.barstow@nokia.com>
- CC: Ian Hickson <ian@hixie.ch>, ext Michael Nordman <michaeln@google.com>, lrbabe@gmail.com, public-webapps <public-webapps@w3.org>
On 4/3/2011 11:20 PM, Ian Hickson wrote: > On Sat, 2 Apr 2011, Arthur Barstow wrote: >> Perhaps subscribers to both lists (Mike Smith, Maciej, Hixie) could >> provide some guidance on which list to use for Offline Web applications >> (again, I'm OK with public-webapps) and which Bugzilla product/component >> to use to file feature requests for Offline Web apps. > If one wants to send e-mail and be guaranteed a response, then the > whatwg@whatwg.org list will be most effective. > > If one wants to discuss the issue within the W3C context, then > public-html@w3.org is the list most appropriate for the current spec text. I'd prefer to see discussion continue on public-webapps . I'm not against cross posting. "Offline" apps are the first and only instance of an 'installable' applications implementation across vendors. At some point, I hope to see some discussion about an installation manifest, for apps which require privileges across origins and/or other enhanced access. This seems to be the direction things are going (from Google's Chrome): <link rel="chrome-application-definition" href="arbitrary-manifest.json"> chrome.app.install() That scheme allows the author to provide a user with a choice, prior to prompting them for permissions. For an application vendor, that's mighty helpful, as some users reject: "Do you want to grant this website access to EVERYTHING--or-it-will-not-work", but accept: "Do you want to grant this website access to this one domain". Google Chrome http://code.google.com/intl/en-US/chrome/apps/docs/no_crx.html chrome.app.install() Mozilla Firefox https://mozillalabs.com/blog/2011/03/first-developer-release-of-web-apps-project/ https://developer.mozilla.org/en/OpenWebApps/The_Manifest navigator.apps.install() Those two vendors are converging. <link rel> seems more appropriate for icons. Apple is taking that route. HTML manifest files seem appropriate for requesting same-origin exceptions, which is part of why I'm keeping an eye on this thread. Those .install() calls are still a good idea, as vendors have their own unique extensions and distribution mechanisms. I don't expect to get background pages / content pages into a web apps spec any time soon. They're the same concept across Firefox/Chrome/Safari, but there are a lot of surrounding details. -Charles
Received on Monday, 4 April 2011 07:07:00 UTC