> > I'm having some problem converting my latex files to HTML. The
> > ps figures in the latex file are cropped in the gif output. The
> > way I included figures in latex file is through the use of \psfig.
> > But the output GIF image is always cropped on the side.
>
> This may be the same problem, I am not sure. I was working with
> encapsulated figures and found that l2h only used what was inside
> the "bounding box" (four numbers appearing after the words
> bounding box in the header of the encapsulated postscript file)
> to create the gif. Sensible really.. but I had a few labels which
> were beyond the bounding box.. they appear in latex OK, but were
> clipped by l2h. I hacked the bounding box.. inspected the results
> with gv (new ghostview) until I was happy, and l2h did fine.
>
Yes, this is the correct way to handle this problem,
when it is due to an incorrect bounding-box.
The purpose of the %%BoundingBox comment in an .eps file
is to convey the areain which the image is to be found.
This is specifically for the benefit of other programs
(i.e. not PostScript interpreters) so that an appropriate
region can be allocated on the page, to hold the image, when it is
eventually drawn; e.g. by a printer.
It is quite unreasonable to expect LaTeX2HTML to decide to ignore
this information, and somehow find out the image size by itself.
Early versions *did* do this, by imaging a complete page for every
image, even if only a single character, at say 10pt size.
This required more than 1Mb of memory, when the final .gif
might be less than 1kb in size.
This meant that huge memory allocations were required for what should
have been easy translations.
Now the %%BoundingBox information *is* used, implicitly via LaTeX.
The \hbox that LaTeX constructs to hold the image is measured,
to determine how much memory to allocate when imaging the graphic.
The decrease in memory requirements, and hence also the time required
to perform a translation, was most dramatic.
So the lesson is, if you use programs that produce incorrect PostScript,
then expect to do some extra work to use their output effectively.
There will be problems using that output with other word-processing
systems too, not just with LaTeX2HTML.
There are some tricks that work within LaTeX2HTML, for some environments;
e.g.
1. if cropped on the right-hand side, then make the $PAPERSIZE larger;
default is $PAPERSIZE='a5';
Large images may require $PAPERSIZE='a4'; or $PAPERSIZE='b4';
or even larger.
This will affect environments which make an image to the full width
of the page, and those that still *do* use the full page size.
2. if cropped on the left, try shifting the image on the page,
for the LaTeX2HTML version only; e.g.
\usepackage{html}
\begin{htmlonly}
\begin{makeimage}
\begin{flushleft}
\end{htmlonly}
\includegraphics{.....}
\begin{htmlonly}
\end{flushleft}
\end{makeimage}
\end{htmlonly}
Hope this helps,
Ross Moore
