Lists Home |
Date Index |
> > No. Every URI _always_ identifies a resource.
>Only if you allow that some things with URI syntax aren't really
>URIs. For example, "http://www.cs.usyrtis.mars/rfc/rfc32767.txt"
>doesn't represent any resource at present: there is no
>University of Syrtis to serve as the authority, and the RFC
>sequence has only reached the 3000s.
>oh right. I suppose the URI has to be 'legal' or 'valid' or
>needs some other such descriptor attached to it.
Arg! I really really really have issues about this (ok, not personal issues
The mere fact that it is possible to create URIs for resources which don't
exist at the time of their creation implies that 'invalid' identifiers are
just as much URIs as 'valid' ones. Take math as a example:
Here's a function, y = 3x + 1
I could come up with a URI for mathematical variables such as
'urn:math:_variable_Name_'. So the identifier for 'y' could be
'urn:math:y', which maps to '3x+1'. The variable 'x' can also be identified
by a uri, 'urn:math:x'. However, 'x' hasn't been defined yet, so:
uri -> representation
urn:math:y -> 3 * urn:math:x + 1
urn:math:x -> undefined
Thus, 'invalid' URIs are useful as well as 'valid' ones when defining a
behavior or function.
Another case for 'invalid' URIs are base URIs. For a document, the base may
be: 'http://www.pitt.edu/~jfcst24/' That URI doesn't point to anything;
there is no representation for it nor is it dereferenceable. Thus, the URI
is not valid. However, the relative URI 'LexX/lexx.html', from the previous
base URI, is dereferenceable and thus valid. This is a another case where
invalid URIs are useful as a base for other relative URIs (so they can be
valid and dereferenced).
"If the path is set in stone, use a sledgehammer"
Join the world’s largest e-mail service with MSN Hotmail.