solyanik, noticed you're only converting the text files instead of HTML. I guess this is because of the amount of traffic the HTML version will cause, i.e. lots of pictures. The results though from the HTML versions are far superior and most etexts are in HTML format now.
Take a look at my
gutlrf.pl script, if you like, which gets the HTML version or falls back to text if HTML isn't available. It also does some cleaning of the Gutenberg HTML file to improve output. My script also uses libprs500 to convert the HTML as that handles the Gutenberg HTML very well now.