Hacker News Reader

There are several ongoing efforts to maintain an archive of SciHub

https://www.reddit.com/r/DataHoarder/comments/8ky647/scihub_...

https://www.vice.com/en/article/pa7jxb/archivists-are-trying...

The paper collection software, if I recall correctly, is actually quite simple and uses APIs provided by publishers. It's the store of credentials that SciHub uses (that are provided by academics and scientists) which truly powers the site

> I'd be much much happier to see the material kept in a more resilient configuration (IPFS, database dumps, ...) so that other entities can back it up.

For IPFS have a look at https://libgen.fun/ and https://freeread.org/ipfs/

Also see the HN thread on those: https://news.ycombinator.com/item?id=25209246

Something like https://collab.ipfscluster.io/, where you could simply decide to become part of the cluster and host data would be nice.

Maybe with some granularity, because I doubt many of us would have enough storage to spare to mirror the whole thing.

It backs up to Library Genesis.

And while the paper collection software isn't FOSS, it's really the idea behind article sourcing that's important.