Le jeudi 04 mars 2010 à 15:51 +0100, Robin Berjon a écrit : > On Dec 10, 2009, at 16:51 , Dominique Hazael-Massieux wrote: > > A quick comment after re-reading WARP at the invitation of Robin to DAP > > [1]: I don’t think the notion of subdomain is well-defined; is w3.org a > > subdomain of .org? is co a subdomain of co.uk? I assume they are in the > > sense of the spec, but if that’s so, it doesn’t match the “street” > > meaning of the word “subdomain”; this matters in particular in section 7 > > (rules for granting access), since this has an impact on how a user > > agent decides to grant access to a network resource. Given that IP > > addresses are allowed, the algorithm to determine if something is a > > subdomain of another domain is as simple as looking to the last dot in > > the authority component. > > That's a fair point. Would referencing RFC 1034 in that section > address your concern? I would rather not have to define subdomain > ourselves but rather reuse what already exists!
Sounds good to me, although I think I would also rephrase somewhat the algorithm, à la: * the URI's scheme component is the same as scheme; and * if subdomains is false or if the URI's host component is not a domain name (as defined in RFC1034), the URI's host component is the same as host; or * if subdomains is true, the URI's host component is either the same as host, or is a subdomain of host (as defined in RFC1034); and * ... Dom