> The deeper concern, for me at 
> least, is that the community hasn't been able to agree on an 
> interoperable means of expressing those extensions.  Shall we 
> all agree to write and distribute them in C#? 

My own view is that users understand these issues. When people ask for help
with XSLT, they will say "I can't use any extensions because the code has to
be portable", or "I can't use vendor extensions, but I can use the
extensions in the EXSLT library because they are implemented in all the
processors I care about". I am prepared to trust users to make these
decisions, and to make users take the consequences if they get them wrong.

A particular argument in favour of extensibility is that it removes the W3C
bottleneck. If W3C defines the extensibility points, then some industry
consortium for (say) geographic information systems can define a set of
extensions and try to get them implemented across a range of schema
processors, and users within that community can then make a decision as to
whether these extensions constitute enough of a "standard" to meet their
needs as a community. Note that the implementations of such extensions don't
have to be portable, only the specifications of the extensions.

In theory of course the current schema specs allow "binary components" and
thus allow vendors to provide the extensibility mechanisms that would enable
such standard libraries to be created by third parties. In practice though
we don't provide enough clues to encourage vendors in this direction.

Michael Kay
http://www.saxonica.com/


From alavinio@g... Wed Dec 12 20:05:19 2007
Received: from wiggum.w