I am also a little hesitant about processing instructions since they
were not under control of the DTD. This meant people could put them in
places that really didn't make sense (like as a child of a para). The
<toc/> and <lot/> elements were under control of the DTD so they were
constrained.
I don't know if the RelaxNG schemas can control processing instructions,
but Norm's email later on solves the problem for us.
Thanks for the thoughts.
Larry Rowland
On Thu, 2006-04-20 at 13:25 +0200, Jirka Kosek wrote:
> Rowland, Larry wrote:
>
> > Since we publish HTML version of all books as well as the PDFs and want
> > them to be as consistent as possible, processing instruction seem less
> > desirable, since it would be necessary to repeat them for the HTML and
> > PDF if the convention of directing the processing instruction at the
> > specific output format is to be maintained.
>
> Processing instructions does not have to be tied to a specific output
> format. You can use some general PIs like <?toc?> <?lot figure?> <?lot
> procedure?> to mark places where ToC or LoT should appear.
>
> I use PIs exactly for this purpose very successfully. I usually use
> customer name as a target for PI to prevent name clashes. So in your
> case, it can be perfectly reasonable to PIs like:
>
> <?hpdoc toc?>
> <?hpdoc lot="figure,procedure"?>
>
> Just my 0.02 PIs ;-)
>
> Jirka
>
