- From: Rob <wlkngowl@unix.asb.com>
- Date: Fri, 25 Apr 1997 05:29:59 +0000
- To: James Aylett <sja20@hermes.cam.ac.uk>, www-html@w3.org
James Aylett <sja20@hermes.cam.ac.uk> wrote in reply to my message: > > A useful feature to add to Cougar would be the use of logical names > > (aliases) for resources using the <link> tag. Links in documents > > would refer to the logical name of the resource, with that name being > > defined (in a shared file) with the actual URL. > > Isn't this part of the idea behind URNs (Universal Resource Names)? I > think this could be achieved better if you used URNs, but specified (via a > LINK element in the HEAD of the HTML document) the specific URN database > or server to use - assuming that is that I've got the ideas right here. Yes and no. My suggestion is for specific aliases for a collection of related resources that are defined and controlled by the document author, meant for ease of maintenance of pages on a site. Combined with the 'scribblings' about the <resource> tag on the W3C site, there's some powerful possibilities. The purpose isn't to provide permanent names of resources, but to simplify maintenance of HTML documents in a way that minimizes broken links and can allow user agents to choose documents based on desired criteria (type, version/date, site, etc.). It allows names to be used across a set of related documents (sections from a book, an institution's web pages). It is not intended to provide permanent and unique global names across the internet. URNs require an infrastructure of servers and a registration authority (or authorities) not yet in place, whereas the ideas I am proposing can be implemented in user-agents. URNs are also not applicable to every resource. (How many personal web pages will be given URNs, for instance?) Comparing link aliases with URNs is like comparing Java with CGI. Rob --- mailto:wlkngowl@unix.asb.com http://www.asb.com/usr/wlkngowl/
Received on Friday, 25 April 1997 05:31:20 UTC