We appreciate the enthusiasm we’ve seen for the simple HTML to EPUB converter we launched last year. To date, the conversions measure in the tens of thousands. And since this uses our core EPUB class to do its work, we will continue to maintain and upgrade it.

The latest improvement is somewhat experimental. Now, by default, the converter splits HTML documents into separate documents. We’re curious to see how this fares. It could raise issues with linking in those documents, but it was needed to address the issue of making EPUB files more friendly with slow processors/connection speeds. If enough of a fuss is put up, we’ll put in an option to revert to the old way of linking to fragments in a single doc.

Among other changes we made, some validation errors have been fixed, all HTML is now processed according to an element whitelist, remote images are supported (use absolute URLS), and various other bugs have been fixed (most notably one that was causing invalid IDs).

Despite all the improvements, we were dismayed to see that converted files have the invariable behavior of crashing Digital Editions, hard. It appears to happen when choosing any section other than the cover. I’ve run them through validation several times, pored over the package files with my own eyeballs and while there are some XHTML namespace issues reported by epubcheck, I can’t find anything serious enough to merit crashing the app as hard as it does. We’re small and have other pressing issues, so we’ve run out of time to troubleshoot it for now. It stands as is. That said, I will happily fix the problem if someone brainy can identify it.