Re: Action item on syntax-based interoperability

On Thu, 2003-10-23 at 13:29, Elliotte Rusty Harold wrote:
> At 11:23 AM -0400 10/23/03, Champion, Mike wrote:
> 
> 
> >Web-related tools such as DOM, SOAP, Xinclude, etc.  How about XSLT?  It is
> >most definitely an important part of the Web as I understand the term, but
> >most definitely not defined at the level of concrete syntax.
> 
> I like XSLT, but is it it a coincidence that XSLT in general is not 
> exchanged on the Web today?

No, it's not a coincidence; it's not even true.

cf

  W3C XSLT Service
  http://www.w3.org/2001/05/xslt

and stuff like
  http://udell.roninhouse.com/bytecols/2002-03-27.html

Now there's a "principle of least power" that argues
against turing-complete languages when you can help it.
But there are a few deployed turing-complete (mobile code)
formats: Java bytecodes, javascript, and XSLT. These are
perhaps less widely deployed than XHTML and the like,
and they have a few additional risks, but architecturally,
they're not any less part of the web.



>  Instead it is processed locally on the 
> server side. In practice, XSLT is as or more unreliable when 
> delivered to clients than JavaScript and DOM. I used to think this 
> was purely because of bad implementations, but now I'm not so sure. 
> Perhaps the problems that plague client side XSLT are endemic to any 
> effort to exchange a data model instead of syntax.
-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 23 October 2003 15:54:13 UTC