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

Dave Pawson dave.pawson at gmail.com
Fri Sep 19 12:41:04 EDT 2008

2008/9/19 David Carlisle <d.p.carlisle at googlemail.com>:
>> With a known and specified repertoire (note that it's not defined in the
>> spec)
>> how can a character be unknown?
> basically the system let's you say which characters are definitely in (the
> kernel)
> which characters are definitely out (those not in the hull)

What assumption are you making there, using the inverse?
" characters are definitely out (those not in the hull && those not in
the kernel)"

 and those that
> you leave as unknown (perhaps because of variants of the repertoire or to
> cope with future versions or whatever)

Chars in 'hull' are 'maybes' aren't they?

3.3 hull
a set of characters that may be in the repertoire

I guess I'm assuming kernel+hull+unknown=repertoire. Perhaps that
isn't a valid assumption.

> so if there is no <hull> in the specification then  the specification
> is definite and characters are either in or not-in,

+1 I'm OK on that part.

 but if there is a <hull>
> specified then the unicode characters are partitioned into three: in unknown
> and not-in
> consider a crepdl schema that just says
>   <char><kernel>a</kernel><hull>b</hull></char>
> so
> a is in
> c is not-in (as is every character other than a and b)
> b is unknown

So the repertoire isn't bounded /  complete then.


Dave Pawson
Docbook FAQ.

More information about the Schematron mailing list