Good news, I have invented an exciting new yak to shave!

Tilde Lowengrimm

Good news, I have invented an exciting new yak to shave!

It would be an understatement to say that I have a lot of ebooks & audiobooks. I think it would be fun to build a virtual card catalog slash index which doesn’t need any extra software to browse — just a nest of symlinks on the file system.

Imagine: you could have all the files named something like Title_Author.m4b, and then another directory with symlinks named Author_Title.m4b. A collection of symlinks organized by Dewey Decimal, another by genre/tag, and yet another directory containing every series, with the symlinks inside numbered so they sort nicely.

Of course, this maze would need to be organized & maintained by software. I’d probably want to store all the metadata in the book files themselves, with the symlinks on disk just being a representation. That would make it relatively easy to rebuild or adjust whenever a new book is added or something changes, and it would be relatively straightforward to support additional labelling system without needing an army of scrivs to constantly prowl about the filesystem. For some reason, my heart tells me that this would be a perfect job for a little collection of shell scripts, even though my better angels warn that way lies madness.

Tilde Lowengrimm

While I’m out here fantasizing about digital library technology, you know what I really want? I want a file format which smushes audio & text editions of a book together with timestamps connecting them. Taking separate files of voice-acted audio & plan text and lining up the timestamps is one of the few non-evil uses I can imagine for modern ML tools. And I simply cannot imagine a better way to read a book than having the audio edition in my ears, but with readable text to jump back to whenever I want to quote something, or see the spelling of a name, or ⌘F for my favorite passage. We have the technology! We could do this!