This is a known problem. Please dig through the mailing list archives.
Gennadiy Tsarenkov has started some work last year to improve JEuclid
itself. He mentioned this problem. Obviously, we will need to extend FOP
just a little so a FOP extension for fo:instream-foreign-object and
fo:external-graphic can report the position of the baseline. Maybe you
can work with Gennadiy to fix this.

On 09.07.2006 22:15:27 Ryan Saunders wrote:
> Hello,
> 
> First of all, I realize that the MathML extension is a work
> in progress, and thus, the current design may be different
> than what I'm looking at (the mathml example that comes
> with the fop  0.92 beta).
> 
> That said, I'm concerned that the approach used by this
> extension is going to have some problems. In the current
> scheme, MathML is converted to SVG by jeuclid, and the
> resulting SVG is then laid out like any other graphic.
> 
> Ignoring, for the moment, the fact that JEuclid seems to
> give incorrect bounds for some equations, there is also
> another problem: the text baseline. The solution of
> converting to SVG and treating it like an image works OK
> for block-display equations, but for inline-display
> equations it does not. Consider a simple mathematical
> expression 'i_0' ('i' with a '0' subscript). The i's
> baseline should be consistent with the text around it,
> while the zero's baseline should be slightly below the main
> text baseline. But this scheme for integrating fop/jeuclid
> would instead put the zero's baseline consistent with the
> surrounding text.
> 
> Perhaps you're already aware of this, but as a fop user who
> would like to see robust equation support, I thought it
> worth pointing out.
> 
> Thanks,
> Ryan


Jeremias Maerki

Reply via email to