Hey wallcraft & bbusybookworm.
Thanks a lot for your quick and instructional answers.
I tried mobihuff first, it gives me the following output:
Copyright (c) 2008 The Dark Reverser <firstname.lastname@example.org>
The output is a single HTML file as wanted, including all of the written content. special characters don't get convertet for some reason and all images are missing. (not only the files itself, but they are not mentioned within the HTML source code, sadly.
Caliber worked fine to open all of the non-DRM test files (linked in my first post) for viewing. But also the files are only between 200 an 900 kB, it takes Caliber several minutes just to open them. (Using an EeePC 1000 HC, 1.6GHz, 2GB RAM, WinXP SP3).
I changed the output format to ePub, just as you suggested and started the conversion. Everything worked fine for 8 or so minutes. Then Caliber crashed. No error messages. I rebooted, gave it a second try and Caliber stopped it's work mentioning some memory problem.
I'll do some googlin'.
Maybe there's a workaround. But I fear the error might lie within the code of the original files?!?
Any more ideas? I'll keep you updated if I'll find out anything new.