Re: New/old pattern syntax, why can't we have both ?

Subject: Re: New/old pattern syntax, why can't we have both ?
From: Paul Prescod <papresco@xxxxxxxxxxxxxxxx>
Date: Mon, 24 Aug 1998 21:43:43 -0500
Pasqualino "Titto" Assini wrote:
> 
> The main, and correct, argument being that as it doesn't conform to XML
> syntax.

That is not at all true. Are you claiming that valid XSL documents are not
valid XML documents? Be careful how you throw around the word "conform."

What you mean is that the new syntax does not express all of its
hierarchical structure through XML elements and attributes. But so what?
We aren't going to reinvent URLs into XML syntax. Nor mime types. Nor HTTP
messages. The line between a query and a URL is pretty vague, as XPointer
demonstrates. 
 
> The new syntax should be seen as a shorthand for the full, verbose, XML
> form.

This makes perfect sense if you come at this from the WWW point of view.
If you come at it from a slightly different point of view, however, you'll
see it differently. Consider:

 * nobody will ever look at the "output" XML of the expansion
 * what you *really* want is a common API between queries and XML
 * actual, textual syntax is irrelevant
 * even as an API, encoding the query in terms of elements and attributes
is needlessly verbose and inefficient

Why not just make an API that can express both queries and XML? While
we're at it, why not throw in other media types like PDF, CGM, MPEG and
XML DTD syntax? Then we could use the best syntax for each data type, but
use a common API for processing it. Check out my SIG.

There is one important reason not to invent such an API: it already
exists. James Clark, one of the two editors of XSL, invented that API
several years ago (more or less...the API is a straightforward application
of the data model). It is even an ISO standard. All languages in the SGML
standard are now described in terms of it. It is breathtakingly simple and
yet powerful and very general.

As I said in my XML Developer's Conference talk: "I've got your freaking
API right here! Use it!"
 
 Paul Prescod  - http://itrc.uwaterloo.ca/~papresco
 ISOGEN International Corporation: "We Transcend Syntax."


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread