Allin Cottrell wrote:
> Thanks, that works fine. One other observation. Your math.xsl
> automatically inserts math delimiters ($ for inline and $$ for
> display) in the PI for xmltex, when it encounters an <alt role="tex">
> element. I think it would be better to let the user insert the
> delimiters as part of the TeX input. Your version fails if the user
> has, say, specified the math with
>
> \begin{eqnarray*} ... \end{eqnarray*}
>
> because then there are double math delimiters in the .fo file, and it
> won't compile.
I'm sorry, I overlooked that your dbtexmath requests user to input math
delimiters. On the other hand in my experience (which can be different
from yours) a majority of equations are simple inline or display ones
without eqnarray and similar environments and thus it is naturel to type
them without delimiters -- delimiters can be automatically inferred from
context -- <inlineequation> or <equation>. This however doesn't solve
eqnarray problem.
There are two solutions:
1. force user to type delimiters on every equation
2. add support for some processing instruction, which will turn-off
automatic insertion of delimiters. E.g.
<alt role="tex">
<?dbtex delims="no">
\begin{eqnarray*}...\end{eqnarray*}
</alt>
I personally prefer 2nd variant, but if your document contains a lot
eqnarrays this is less suitable for you.
We should setlle on some common practise definitely as it isn't good
when DSSSL and XSL implementations aren't 100% compatible. So question
is: variant 1 or 2?
Jirka
--
-----------------------------------------------------------------
Jirka Kosek
e-mail: [EMAIL PROTECTED]
http://www.kosek.cz
