- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Thu, 7 Aug 2008 17:28:27 -0700
- To: "Schleiff, Marty" <marty.schleiff@boeing.com>
- Cc: <www-tag@w3.org>
On Aug 7, 2008, at 3:08 PM, Schleiff, Marty wrote: > I haven't seen any messages complaining about a URI scheme for > widget:. I've seen plenty of messages complaining about a URI > scheme for xri:. Defining a special addressing scheme for widget is a bad idea. As has been pointed out many times over the years, packaging multiple representations into a single archive cannot rely on the ability to rewrite references within the content of individual parts because the individual parts may be cryptographically signed before the package is created. The only solution that is known to work for all of these scenarios is to embed a catalog part that tells the local resolution process which parts of the archive correspond to which URI references, thereby allowing any URI reference to be used to refer to any part. If a part is created without its own URI, the cid scheme is the recommended choice for minting new URIs within a package. If a packaging format needs to control references to outside content, it should do so as part of the local resolution algorithm, not by minting an application-specific identifier. This should apply to all such bundling mechanisms, whether they be web archives, zip containers, OSGi bundles, or waf widgets. Cheers, Roy T. Fielding <http://roy.gbiv.com/> Chief Scientist, Day Software <http://www.day.com/>
Received on Friday, 8 August 2008 00:29:05 UTC