I also thought this initially, but actually the time-span group is browsed in sequential chunks, so we can lazy loading them.
The only cons of this i approach seems that you cannot have the real total count for every user of all the published events.
Wait, I'm confused. I thought you were fetching all follows from last X hours into a map in cache and displaying that, with lazy-loading in the rendering step, when someone clicks on an entry.
How does this work?
This is really good, by the way. I'll host it on my https://jumble.imwald.eu .
An AUTH aggregator like aggr.nostr.land really shines with this view nostr:npub12262qa4uhw7u8gdwlgmntqtv7aye8vdcmvszkqwgs0zchel6mz7s6cgrkj as you can deliver big data fast without triggering rate-limiting.
This actually helps frequent posters, too, as you can filter them out, temporarily, instead of muting them. Even if you only used this view, they'd reappear during times when they were quieter.
