[Schematron] Help sought: implementation of Character Repertoire in XSLT2 for embedding in schematron

[Dave Pawson]

2008/9/19 David Carlisle <d.p.carlisle at googlemail.com>:

>> The negation (not-in) seems a straight-forwards Not (match (., 'regex')).
>
> No. not-in isn't the negation of in, it's a three valued logic, in not-in
> and unknown.
> You need to evaluate one test to see if is definitely in the repertoire,
> another similar one to see if it is definitely not-in and then if it is
> neither of those things it is unknown.
>
> However if having done the test for in you do a quick test and check that
> there are no kernel or ref elements in the crepdl file, then I think it is
> the case that unknown can not be the outcome, so you don't need to test for
> not-in, the not-in property will in that case just be the negation of the in
> property.


I'm quite confused about the Venn diagram for this.

given character c, repertoire R, from the 0807 version:

Section 4 provides notation.

in(x, A): character x is in repertoire A
not-in(x, A): character x is not in repertoire A
unknown(x, A): it is unknown whether character x is in repertoire A

With a known and specified repertoire (note that it's not defined in the spec)
how can a character be unknown?

7.1 has

Given a repertoire description A, either in(x, A), not-in(x, A), or
unknown(x, A) holds.

Does that mean A | B | C, or a selection of one and only one from the 3?


Then given the semantics, there seems no rationale for the 'hull' set?


I really don't understand this document.


regards DaveP






-- 
Dave Pawson
XSLT XSL-FO FAQ.
Docbook FAQ.
http://www.dpawson.co.uk