Silly Question - do we need to explicitly say what to do with illegal URIs?

There are several attributes that take URI or URIorSafeCURIE as their 
datatype (@href, @src etc).  Do we need to explicitly indicate what 
should happen if a resulting URI is illegal?  Examples:

   1. Multiple URIs in an attribute that takes a single URI (e.g., @href).
   2. URIs that contain illegal characters.
   3. Prefix mappings that contain illegal characters.

I don't know - personally I am not a fan of attempting to identify every 
pathological case and define error behavior.  In the XHTML 
specifications we generally say that we are defining the behavior when 
the spec is used correctly, and otherwise the behavior is unspecified. 

What brings this up is I was updating my parser, and noticed that it was 
possible to define an @href that contained multiple URIs.  I was about 
to change my code so it would magically deal with it, and I thought 
"that's a stupid idea.  I don't even know what the right thing to do 
here is."  Now I silently ignore it because it is not a valid URI 
"obtained according to the section on CURIE and URI Processing".  But 
that section doesn't actually say anything about what a valid URI is - 
just what a valid CURIE is.

-- 
Shane P. McCarron                          Phone: +1 763 786-8160 x120
Managing Director                            Fax: +1 763 786-8180
ApTest Minnesota                            Inet: shane@aptest.com

Received on Friday, 9 May 2008 16:04:17 UTC