06. Semantics [ref. JP32 034]

[AndrewSales]

This clause lacks the most important part: a processing model for (1) choosing phases, patterns, and rules, (2) establishing rule contexts by evaluating context expressions, (3) evaluating assert and report elements in selected rules, and (4) compiling the SVRL report. At present, this clause defines ancillary mechanisms only.

[rjelliffe]

The initial version of Schematron specified the action of validation, phases etc using predicate logic. For the Rev 2 draft, the Japanese national body objected that it was not correct, but could not point to how it was correct, or provide corrected text. I could find no-one credible who understood the logic notation and who was interested in contributing (I teeed up a local Computer Science professor at the #1 Uni here, but they lost interest). At which stage, depression set in that I was in a Catch 22. IIRC, I followed the Japanese request to move that text to a non-normative note. But I still believe it was a mistake: the notional validation operation of Schematron is a few nested case statements, and the standard would be better if it gave a semi-formal statement (using pseudo-code or logic).

…

On Sat, Apr 22, 2023 at 11:21 PM Andrew Sales ***@***.***> wrote: This clause lacks the most important part: a processing model for (1) choosing phases, patterns, and rules, (2) establishing rule contexts by evaluating context expressions, (3) evaluating assert and report elements in selected rules, and (4) compiling the SVRL report. At present, this clause defines ancillary mechanisms only. — Reply to this email directly, view it on GitHub <#56>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AF65KKKWTEZDOCX7WYURKBTXCPLPJANCNFSM6AAAAAAXH232GI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[AndrewSales]

Added an informative sub-clause to Clause 6.