- From: Phil Archer <phil.archer@icra.org>
- Date: Wed, 15 Dec 2004 14:24:25 -0000
- To: "Quatro Public list" <public-quatro@w3.org>
Dear all, I'm working on the new ICRA label generator and, since it asks for user input, it's 99% error checking of course. I'm writing the code to extract the domain info which the draft rule spec calls for. This is made easier by Perl's URI module that has a nice little method that returns the host for a given URL. OK. So: http://www.example.org returns www.example.org. So far so good. I can strip off the www and use that. Then we have http://subdomain.example.org. The domain is still example.org but the host is now subdomain.example.org. OK, rather than writing a rule to match ".*" I can now write it to match ".*subdomain.example.org.*". This is going to be important for people who have homepages on big ISPs whose websites have addressed like mydog.btyahoo.com But extracting a domain from a host is not always easy. It's OK for TLDs like .org and .com - you just take the bits either side of the last "." - kind of breaks down with example.co.uk though. So... as discussed earlier, the restriction is important. We need that to limit the scope of a label that matches ".*" but on reflection, can it be host rather than domain? This takes care of those personal websites and a heap of other stuff. If we have: <rule:hostRestriction>example.org</rule:hostRestriction> Then "matches .*" should still match all subdomains of example.org. But we can also have: <rule:hostRestriction>subdomain.example.org</rule:hostRestriction> Which would only match against that host or its subdomains. In essence, the value of the hostRestriction must match the right hand side of the host within the URI under test. OK? Phil.
Received on Wednesday, 15 December 2004 14:29:22 UTC