Re: [xml-dev] [Summary] Characterization of Schematron - Usage and Featu

["Peter Hunsberger" <peter.hunsberger@gmail.com>]

On 1/25/07, bryan rasmussen <rasmussen.bryan@gmail.com> wrote:
> I agree it will be easier in iso schematron , because of let
> basically, and if you're bound to xpath 2.0 as you have for and
> conditional expression. I don't feel comfortable doing it for public
> projects yet though, given the newness of all standards used in such a
> scenario.
>

We've been using Schematron with XSLT 2.0 and xPath 2.0 in a high
visibility production system since shortly after Saxon 8.0 came out.
We have a set of Web pages where we need to loop over an array of
values looking for co-constraints (as well as the usual value
constraint stuff).  We use an intermediate XSLT that takes an ordinary
Schematron rule (generated from an front end administration editor
that our business analysts use), recognises that the particular use
case needs to run over an array and re-writes the xPath appropriately.
 We could have just picked some large upper bound (say 100 rows) and
automatically written rules for every row but the XSLT and xPath 2.0
solution is a lot simpler.

As to your other note on the size of xPath expressions growing, most
definitely yes.  We already see this, since our business analysts are
using an editor that hides most of the xPath they seem to have no
problem writing rules that run for many lines.  Simple for them, but a
real pain when something breaks and the developers have to figure out
why a particular set of validation rules isn't yielding the results
they expect.
-- 
Peter Hunsberger