Hi List!
Here is the exact URL to what I mentioned yesterday.
Please go and take a look at how the scanned image has
been combined with the "raw" OCR output:

http://www.lysator.liu.se/runeberg/gudasaga/0053.html

If we are going to do something similar on this list,
then the way they've done it as Linköping University
in Sweden, might be an instructive example. The image
is just 49.5 K - that is per page of text. The picture
format is something called *.pike, which is new to
me. But Netscape can read it without any problems!

Best regards,

Keth