Re: [relax-ng-comment] Mixing patterns and (empty) tags

[Daniel Prager <danielp@deakin.edu.au>]

Hi Eddie

Eddie Robertsson wrote:
>  There might be one more option if the validation time of the application
> isn't critical which is to embedd Schematron rules in your RELAX-NG
> schema to check that the unrestricted text content match your defined
> constraints.

[snip example]

This sounds promising ...

> In Schematron you have access to all the String and Number fucntions in
> XPath so can make fairly complex assertions about the text even if you
> can't use regular expressions

... except that in my case I really do need regular expressions!  Oh well.

Actually, I've held off going the hybrid Schematron + [RNG or XSD] route for
a while now, but may well look into it soon.

The particular work I'm doing has been interesting in that on a couple of
occasions
so far I have come upon real-life examples that strain the expressivity of the
schema validation languages.  To what extent is completeness a design objective
of RELAX NG?

Cheers

Daniel

> Hi Daniel,
>
> >Thanks for your explanations and suggestions.
> >
> >It seems that I currently have two options:
> >
> >1. Unsafe: Use unrestricted text and check the text elsewhere
> >   (as originally suggested my Mike).
> >
> There might be one more option if the validation time of the application
> isn't critical which is to embedd Schematron rules in your RELAX-NG
> schema to check that the unrestricted text content match your defined
> constraints. Here is an example of how this could look for the example
> provided by Mike:
>
> <?xml version="1.0" encoding="UTF-8"?>
> <element name="example" xmlns="http://relaxng.org/ns/structure/1.0";>
>     <sch:pattern name="Check the text content"
> xmlns:sch="http://www.ascc.net/xml/schematron";>
>         <sch:rule context="example">
>             <sch:assert test="text()[1] = 'The quick '" >The first text
> node must be 'The quick '.</sch:assert>
>             <sch:assert test="text()[2] = ' brown fox '" >The second
> text node must be ' brown fox '.</sch:assert>
>             <sch:assert test="text()[3] = ' jumped over the lazy dog 5
> times.'">The third text node must be ' jumped over the lazy dog 5
> times.'.</sch:assert>
>             <sch:assert test="normalize-space(.) = 'The quick brown fox
> jumped over the lazy dog 5 times.'">The whole text must be 'The quick
> brown fox jumped over the lazy dog 5 times.'.</sch:assert>
>         </sch:rule>
>     </sch:pattern>
>     <interleave>
>         <text/>
>             <element name="a"><empty/></element>
>             <element name="b"><empty/></element>
>             <element name="c"><empty/></element>
>     </interleave>
> </element>
>
> As you can see you can make assertions about each text node separataly
> or as a whole. In Schematron you have access to all the String and
> Number fucntions in XPath so can make fairly complex assertions about
> the text even if you can't use regular expressions. For more info on how
> to embedd Schematron rules in RELAX-NG schemas (or W3C XML Schema) see [1].
>
> Cheers,
> /Eddie
>
> [1] http://www.topologi.com/public/Schtrn_XSD/Paper.html
>
> >
> >2. Inelegant: Wrap the text bits in their own tags
> >    (as more recently suggested by Mike).
> >
> >I'll probably be going with the former for the moment (since
> >I can check non-schema constraints elsewhere if needed).
> >
> >However, I'm curious as to whether this limitation could be lifted.
> >It seems to me that this might be done either using the <interleave>
> >idiom, or without by requiring each text fragment to form a complete
> >pattern.
> >
> >Opinions?
> >
> >Thanks again
> >
> >Daniel
> >
> >--
> >Dr Daniel Prager
> >Research Fellow & Consultant
> >Deakin University & Freelance
> >
> >Phone: +61 3 9251 7484
> >Web: www.deakin.edu.au/~danielp
> >
> >
> >
> >----------------------------------------------------------------
> >To subscribe or unsubscribe from this elist use the subscription
> >manager: <http://lists.oasis-open.org/ob/adm.pl>
> >
> >

--
Dr Daniel Prager
Research Fellow & Consultant
Deakin University & Freelance

Phone: +61 3 9251 7484
Web: www.deakin.edu.au/~danielp