>perpetual accumulation of the unread count
In practice, is that an issue? you could just ignore that number, or mark all as read when you are done reading. It looks like the example you gave in link 2 has some unread counts perpetually accumulating, in fact, unless I misunderstand.
sfeed_html just outputs a flat html file, I know it looks like unread counts, but it's just showing a count of entries with a published date within the last 24 hours of when the page was generated. It's just an easy way to draw attention to "new" entries.
An unread count feels to me like another todo list. I like not having it.
The reason I couldn't stick with other feed aggregators in the past is because of the perpetual accumulation of the unread count. The output of sfeed_html is just links [2]. No unread counts. I can remember whether I've read something or not.
Just stick it on a server, run as a cronjob, and you have your new browser homepage. And if you want to move your "database" it's just a directory of tsv files. Brilliant.
[1] https://codemadness.org/sfeed-simple-feed-parser.html
[2] https://codemadness.org/downloads/sfeed/html/feeds.html