I've spent the last three days working to fix the ePubs and AZW3s of Yesterday's Tomorrows. I had just ran it through Calibre and did a quick check, noting that the table of contents didn't display anything.
It took a lot of research and learning to fix the ToC, and while doing so discovered something even worse - some of the illustrations were covering up the text. Damn!
Trying to figure out the ToC I tried several things. One was installing the Write2epub extension to Open Office.
It really sucked, especially with this book. It had some ugly sans-serif typeface, and there were huge swaths large and bolded that I never told it to do. And there was still no table of contents.
While googling and reading and finding out that e'books were mostly based on HTML5, XML and a few other things, I got a little disheartened. This was going to take forever, because I had a lot I had to learn.
I ran across Google's e'book editor "Sigil" and installed it. I have no idea if it's any good, because there's no documentation and I can't make heads or tails out of it.
So I went back to Calibre and studied it some more, educated a little but not much by the internet, and saw a long string with an "and" in it, "h1 and h2" and recognised this from HTML and the rest of the garbage from programming for thirty years. Stupid Calibre was telling it to make everything part of the table!
It took a bit of trial and error to get the right parenthesis and brackets in the right spots that the conversion wouldn't crash with an error, but I finally got a working table of contents.
Now to address the obscured text. That took quite a bit of head scratching as well.
I finally just decided to make the input make the output behave, rather than trying to tweak the output itself. What finally worked was to load the offending images in GIMP and add a white space where it was covering the text. That worked.
So if you've already downloaded one of the e'books, you should delete them and download the new version.
I think I'll take the day off tomorrow.