One was a Fever clone that had a number of strategies for doing parallel fetching:
- https://github.com/rcarmo/bottle-fever
Andrea Peltrin took that and evolved it into Coldsweat: https://github.com/passiomatic/coldsweat - which I recommend if you want a web UI.
I did a number of other things, but eventually went back to what I used _before_ Google Reader: e-mail.
I was one of the contributors for http://newspipe.sourceforge.net/, and after getting bottle-fever going I decided to investigate the state of the art and did a quick fork of rss2email that injected messages into an IMAP store instead of sending them via SMTP, to avoid spam traps.
It was a quick hack, but it allowed me to read feeds using any mobile IMAP client, and a friend eventually did a Go version, which I've also tweaked to my liking:
- https://github.com/rcarmo/rss2imap (Python) - https://github.com/rcarmo/go-rss2imap (Go)
Any of the above are likely to save people a fair amount of time (do bear in mind that the Python version was a hack atop code that was written by Aaron Swartz a decade ago, and it shows its age).
These days I ended up going back to Feedly, simply because I have to use Windows, the Web UI is good enough and there are lots of good clients for the platforms I use (NextGen, Reeder, etc.)
Plus I realised that trying to archive stuff from hundreds of feeds was somewhat pointless -- the stuff I really want to keep around goes into Pocket or OneNote, and that's that.
Edit: Also, here are some notes from 2008 on Bayesian classification and its effectiveness: http://taoofmac.com/space/blog/2008/01/27/2203#an-update-on-...