Re: A Modest Proposal for the Re-inflation of XPath to an XM

Vun Kannon, David <dvunkannon@xxxxxxxx> wrote:
> A Modest Proposal for the Re-inflation of XPath to an XML Syntax
> [...]

Nice idea. I doubt whether it would be acceptable as part of the XSLT
draft - and more to the point, whether it _should_ be included.

Instead I would consider defining XSLT in two levels. The current draft is
the "machine language" level. It might even make sense to omit some things
from it (attribute sets? modes?).

The other level is a "high level" one. It would include anything omitted
from the lower level, loops (side effect free!), something along the lines
of your proposal, and whatever other constructs which would turn out to be
common/useful for XSLT stylesheet writing.

The transformation from the "high level" to the "low level" would be
specified, obviously, in XSLT. The "low level" version of this sheet would
serve as the formal specification of the high level language. Probably this
would be generated from a high level sheet, though, after a period of
bootstrapping.

This has advantages for both XSLT processor implementors and for XSLT style
sheet writers. As things stand, there's an inherent conflict between the
need to keep the language simple and efficiently implementable and the need
to make it friendly and support good software engineering practices.

Of course, all this can wait until XSLT 1.0 rolls out the door. But it would
make a pretty good direction for XSLT 2.0, IMVHO.

Have fun,

    Oren Ben-Kiki


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