Then again, to download anything (including screenshots and other app info, let alone the app itself) you need a server connection so it's only mildly useful to not have to ask a server. And the search algorithm is... nonexistent I think is the most correct word to use. You get what you ask for. The top three hits for "camera" are Telegram and two identical looking SIP clients, of course. (There's not being influenced by algorithms and there's this.) I wonder if te project would accept a better algo if I'd contribute one; I've made one before at work for a not dissimilar dataset so I have some notion of what works, though it's not rocket science anyway.
That said, apk diffs seem important as well, some updates are significantly large (more than 100MB) and see frequent updates. Although in that case it might need to cache apks which requires space as well (I would opt in, though!).
This isn't something like a news reader.
I think homebrew moved away from repository approach too.
Your assumption on how it works is flawed to begin with. Check the comment from one of the original authors https://news.ycombinator.com/item?id=35003844