[sc34wg3] XTM 2.0 - C14N of values with datatype != xsd:anyType

[Lars Marius Garshol]

* Lars Heuer
>
> Why does the XTM 2.0 draft mentions only the C14N of XML? Doesn't it
> make sense to point out that other datatypes should be canonicalized
> as well? Why has XML such an outstanding position?

We've gone back and forth a little on this, but at the moment we're  
not actually canonicalizing any data type representations for the  
sake of duplicate suppression. XML gets special treatment for two  
reasons:

  - To ensure that there is a definite string representation of
    the embedded XML. Without canonicalization we wouldn't have this,  
since
    in most cases the original raw XML markup will not be available to
    implementations (parsers usually don't expose it).

  - To ensure that all necessary namespace declarations are preserved in
    a consistent way.

We could do this without using the XML C14n spec, but when a spec  
that does exactly what we want exists it seems silly not to use it.

> It wouldn't make sense to put a list of datatypes into the spec, but a
> hint that other values require c14n makes sense, doesn't it?

At the moment this is not required. In fact, if you are to follow the  
spec to the letter you *shouldn't* canonicalize. However, this is one  
area where I expect that some implementation experience will  
eventually inform the spec. Ie: if in the 5-year review we find that  
most implementations actually do the c14n, well, then we know what to  
do...

--
Lars Marius Garshol, Ontopian               http://www.ontopia.net
+47 98 21 55 50                             http://www.garshol.priv.no