I haven't used Calibre for years (it was annoyingly slow on a low grade laptop), but I've dealt with many epub files, and I think Calibre does a bad job when mo

I haven't used Calibre for years (it was annoyingly slow on a low grade laptop), but I've dealt with many epub files, and I think Calibre does a bad job when modifying them. It's easy to identify an epub that went through Calibre: there's usually a lot of junk metatada in the file. Not just a "contributor" tag or a custom "modified" timestamp, but dozens of empty json structures embedded in xml. And a (potentially empty) bookmark file. But what surprised me is that Calibre accepts to write metadata that is obviously wrong. For instance, a simple heuristic could detect when the user has confused the full name "Jane Doe" with its file-as form "Doe, Jane" and ask for confirmation. And after seeing many dates at "0101-01-01 ...", I suspect that was the default value in some version of Calibre. I'm glad Calibre exists and that many people like it, but I hope a better software will emerge: safer, saner, faster, maybe prettier and easier to install.