RE: Possible new issue on interpretation of relative URI actors

>I would think it is absolutize and canonicalize.  URI's can 
>have things like escaped characters as well.  Further, XML 
>allows characters that are illegal in URIs.  We need to say 
>something about illegal URI characters in SOAP roles.

The current text about URIs in SOAP suggests that we don't say anything
about how URIs compare. The reason is that URI comparison is a function
of the URI scheme and as SOAP doesn't say anything about specific URI
schemes, it can say nothing about how these may compare other than
referring to RFC 2396.

>XLink section 5.4[1] describes one mechanism for computing an 
>absolute and canonical URI and handles illegal characters.

For the sake of stability and interoperability of the Web, I find it
disturbing that so many specs seem to have their own idea about how to
deal with URIs. I am wondering why it was not considered to link to [2]
for encoding rules? 

Henrik

>[1] http://www.w3.org/TR/xlink/#link-locators

[2] http://www.ietf.org/rfc/rfc2718.txt