Re: DTD from extended RNC

On 21 May 2018 at 14:42, maxwell <[email protected]> wrote:
> On 2018-05-21 08:17, Dave Pawson wrote:
>>
>>    Anything you can do with a DTD you can do with rng/rnc.
>
>
> But surely not in the opposite direction, correct?  Which is presumably why
> trang is complaining.

Yes, rng is more recent / comprehensive than DTD's.

>
> FWIW, we had an analogous problem: we want to produce PDFs from XML
> documents about linguistics.  We chose DocBook as the base, and got rid of a
> ton of things in the RNG/RNC that were irrelevant to our use case, then
> added a few that we needed for linguistics.  The authors of our linguistics
> books use XMLmind (XXE), which I see someone else suggested.  But you prefer
> to stay with emacs, which is understandable.  I myself use a programmer's
> editor, but it doesn't do any schema validation (it does check for
> well-formed XML); so I post-validate against the schema.

Likewise. James Clarks original was syntax directed. I can't get that to work
with the emacs version.
  If you want comprehensive syntax directed, oXygen is (IMHO) the best
way to go. Do your transforms within the editor too.

>
> BTW, we looked into the usual XML-to-PDF via XSL stylesheets, and decided
> they were not sufficient for our needs (right-to-left scripts embedded in
> left-to-right text, complex scripts,...).  Instead we use dblatex (not to be
> confused with the older db2latex), which as the name suggests produces a
> latex format, then use (xe)latex to produce the PDF.  This of course also
> needed modifications for our linguistic use case.

There is one xsl-fo to PDF that might work for you, but it is pricy.

HTH





-- 
Dave Pawson
XSLT XSL-FO FAQ.
Docbook FAQ.