Is there a bug in SVRL generation?

[JoergWaltherOdette]

The ISO Schema for svrl specifies:

sources: https://github.com/esacinc/sdcct/blob/master/sdcct-core/src/main/resources/META-INF/sdcct/schema/schematron/svrl.xsd https://github.com/Schematron/schema/blob/main/svrl.rnc The xml generated by the schematron artefacts produce: i.e. documents versus document What is the correct version?

[phax]

It depends on the ISO Schematron version you referring to. Details follow later....

[phax]

So, here we go:

• ISO/IEC 19757-3:2006(E) Did not define this attribute at all:

• ISO/IEC 19757-3:2016(E) Did define it as "documents":

• ISO Schematron XSLT:

  2009-02-19
    * RJ add experimental non-standard attribute active-pattern/@document which says which
    document is being validated from that point to the next similar. This is to cope with the
    experimental multi-document validation in the XSLT2 skeleton.

As the latest version was published in 2010, they never updated it. See https://github.com/Schematron/schematron/blob/master/trunk/schematron/code/iso_svrl_for_xslt2.xsl#L551 for the code of the last version

• And finally SchXslt does it according to Schematron 2016:

So the question is: remain on the "de facto standard" Schematron XSLT and use document or built something that is correct in regards to the latest ISO standard and use documents? In my Java Schematron implementation ph-schematron I decided to stick with document until further notice (see e.g. https://github.com/phax/ph-schematron/issues/149) but it might be time to change this. My compromise will be, to emit document and documents together :)

[JoergWaltherOdette]

Hi Philip,

Thanks for the analysis and thorough answer. Your proposal sounds very reasonable to me.

Best regards / beste Grüße Jörg