Re: Poll: Relative URIs and Strings in xmlns attributes

At 00/09/20 11:49 -0400, Joseph M. Reagle Jr. wrote:
>At 15:34 9/20/2000 +0100, merlin wrote:
>>I could accept #1 if the spec provided a precise but
>>simple pseudocode algorithm for determining whether
>>a URI is relative or not. Are URIs always URLs?
>
>This is an interesting point I hadn't considered, that #1 requires you to 
>add code to determine if it is an absolute URI. The pseudo-code should be 
>based on the absolute URI definition.

Okay so far.


>"To be defined, the name space node value must meet the following 
>production rule from [URI, section 3]:
>    <scheme>://<scheme-specific-part>
>A namespace node value which does not meet this production rule is 
>undefined -- and it consequently can not be canonicalized by this 
>specification."

Not exactly.

>[URI] http://www.ietf.org/rfc/rfc2396.txt

Mailto, mid, cid,... are all not relative, but don't meat the above
production. These schemes are called opaque schemes. With exception
of some backwards compatibility explanations, an URI is absolute
if it meets <scheme>:<whatever>.

Regards,   Martin.

Received on Thursday, 21 September 2000 05:41:45 UTC