Re: Co-validating stylesheets?

["Oren Ben-Kiki" <oren@xxxxxxxxxxxxx>]

disco <disco@xxxxxxxxxxxxxxx> wrote:
> What I'm rusty on is the notion of Turing-completeness. I might be
> demonstrating my own ignorance here, but I'm wondering if somebody could
> provide a simple example of a stylesheet that's just impossible to figure
> out whether it co-validates.

Forget about Turing for a moment. Consider the following stylesheet
structure:

- Compute a Boolean attribute of the input tree.
- Use a certain template only if the attribute holds.
- The template code produces correct results only if the attribute holds.

The problem is that the checker needs to convert from "the attribute
holds" - which it can figure out - to what this means in terms of the input
document. This is difficult (impossible in the general case).

For example, the attribute might be "date X is not a Saturday", using the
standard Gregorian calendar (not a pretty code, but doable directly in
XSLT). The invoked template reads in formatted calendars - say as tables,
with one  row per week. It assumes that the date "X + 1" is the next cell of
the same row as the date "X". A bit far fetched, admittedly, but you get the
idea.

I have no idea how common such a scenario is in practice. What's certain is
that an XSLT stylesheet validation tool would make a good thesis for someone
:-)

Have fun,

    Oren Ben-Kiki



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