- From: Marcos Caceres <w3c@marcosc.com>
- Date: Mon, 4 Nov 2013 12:17:13 +0000
- To: Anne van Kesteren <annevk@annevk.nl>
- Cc: Noah Mendelsohn <nrm@arcanedomain.com>, "www-tag@w3.org" <www-tag@w3.org>
On Monday, November 4, 2013 at 12:11 PM, Anne van Kesteren wrote: > On Fri, Nov 1, 2013 at 8:01 PM, Marcos Caceres <w3c@marcosc.com (mailto:w3c@marcosc.com)> wrote: > > Only architectural question I had right now is if this should be layered on top of fetch [1]. I guess it depends if the class of application that the app:// uri scheme is used with is considered part of the web platform. > > > > [1] http://fetch.spec.whatwg.org/ > > The web platform works on URLs that are universal. So e.g. http/https > works due to DNS. A still-somewhat-theoretical p2p might work due to a > universal content identifier. The application you are referring to > however cannot be addressed by either of these. Rather, to get it you > need to go through a "store" and download a "trusted" blob. That's > very different from what the web platform stands for. Agree. > Or in other words, you are designing a proprietary walled garden > system around web technology. And that's fine, but the web platform > should aspire to higher goals. Agree also. Maybe we should publish this as a Note? That is, so people can still have some interoperable spec that they can use for these situations, but not make it something the W3C “recommends”. Also, my question was about layering of app:// on fetch… not so much about the evils that app:// enables:) -- Marcos Caceres
Received on Monday, 4 November 2013 12:17:44 UTC