[l2h] testing l2h992-beta5 ...

[Ross Moore]

> Hi!
> 
> > Your image is in .pcx format, inserted into the document
> > via \special{em:graph kurven.pcx}
> > 
> > This \special command is specific to one dvi-driver only
> >  (em-TeX, I'd expect) and is certainly *not* a standard
> > TeX or LaTeX command. It cannot work with a Web2C-based
> >  teTeX distribution, as are most Unix-TeX installations.
> 
> dvips understands this special command. I have several other
> pcx-pictures,  which are correctly interpreted and included into the
> html-files as gifs.

Yes, it knows how to allow a .pcx file to be displayed.
However it does *not* know how large it is,
or any other properties concerning its contents.

This is the real cause of your difficulties.

One way to work around this is to:

\usepackage{graphics}

and then put the {picture} inside a  \framebox 
Then the images created by LaTeX2HTML will have size
governed by the size of the resulting frame.
 
 
> As I said, images.tex is okay, so is images.ps if you copy the pcx
> picture into the right directory. 

Yes; either a copy, or a symbolic link will do.


> > Note that LaTeX2HTML does *not* know to look within the
> > parent directory for files named in \special commands,
> > since there is no standard for such commands being used
> > in this way.
> 
> It does not know when images.ps is created manually, but latex2html does
> know when dvips is called to build.

OK; we must have programmed that to be smarter than I recall. ;-)


> 
> Anyway, I just found the reason. The picture is some kind of
> combination. It uses the pcx picture and adds some normal text with the
> put command. If you remove the puts, the picture is integrated into the
> html. This worked in previous versions. What is wrong here?

Yes. In the combination,  dvips -E  determines the size of the image from
the location of font characters and the lines from \\(h|v|)rule  commands.
Thus it gives you a rectangle enclosing just the labels.
If you \put a '.' in the top corner (left or right) then you would
get the full rectangle that you desire.

When the \special is the *only* thing on the page, then  dvips  returns
a full page. It is left to the cropping utility to reduce to a minimum
rectangle enclosing all of the ink.


If you have lots of such images, I'd advise that you use Ghostscript
to read them and write out a .eps version.
Now LaTeX's \includegraphics command can read the %%BoundingBox 
comment, and both LaTeX and LaTeX2HTML will be able to let you
do more things with your images.

As for putting labels on the images, have a look at what can be
done using the Xy-pic macros, and WARMreader.sty :

  http://www-texdev.mpce.mq.edu.au/TUG/WARM/WARMarticle/
  http://www-texdev.mpce.mq.edu.au/WARM/


> Greetings
>       Andreas

Hope this helps,

	Ross