Ah, cool, that makes good sense. Yeah, it'll be a good learning process, as Django does handle so many of those things basically like magic and you never really need to learn what's going on under the hood. Good luck!
Out of interest, why are you avoiding using a framework? I use Django literally every day for web dev, so I'm curious as to what your site requirements are like.
I'm still using it because old-dot-reddit-dot-com still works, and until it doesn't, I probably will. That said, I'd rather the fediverse thrive than the increasingly corporate-beholden reddit does, so I'll favour what sparse engagement I make to a lemmy instance first.
I think what's hardest to replace from reddit is the absolutely monstrous archive of posts and discussions, which seems to be a bit of a two-edged sword for them (if the official statements are to be believed) - it costs a tonne in hosting, but makes them the most relevant source for real human discourse. This needs to be handled better, and ideally I'd want to see:
- Some sort of archive-dot-reddit-dot-com. Minimal, flat html, ideally anonymised as much as computer-ly possible to help with the inevitable privacy issues this would raise.
- Some sort of mobile-dot-old-dot-reddit-dot-com, as they seem incapable of making an app without bloaty (both visual and bandwidth wise) "features". Call me a boomer, but if I can do something without a specific app, I would rather do it that way.
- Separate i-dot-reddit-dot-com and v-dot-reddit-dot-com into different companies from the main reddit, reddit should be link aggregation and discussion, content hosting seems like a costly thing to try and monopolise.
- If it really costs so much to run the APIs, I'd rather see more user-based rate limiting than price gouging to discourage bad actors. I do not think that is why they are price gouging, but am trying to assume good faith on their part for discussions' sake.
I know I'm an idiot, and some of these are possibly already done and I just haven't looked hard enough, probably some are impossible for obvious reasons I haven't seen. Though even if reddit as a company turned around and tried to become a curator of the discussions it holds rather than milk it's current audience dry with ads, I'd still rather see lemmy out-compete it. Protocol > Platform.
You're still right, though - talking about closest planet on average isn't very useful, because it's always going to be the closest planet to the sun. Asking "what planet can get closest to some [Planet]" is more interesting and enlightening.