Kind of the opposite, sci-hub isn't really curated at all. Its goal is to have a PDF for everything that has a DOI, more or less. And the primary way of retrieving papers is to already know the exact paper you're looking for, found via some other search method like Google Scholar. It's not at all focused on categorizing or browsing papers, just slurping all of them into a big repository.

> Kind of the opposite, sci-hub isn't really curated at all. Its goal is to have a PDF for everything that has a DOI, more or less.

Side-stepping the politics and legality for a moment, I find it interesting how this is a very natural example of the fault-tolerant design of networks.

A 'DOI' is a resource locator. There's a 'fault' in the network in that some people cannot locate the resource when providing the DOI. Sci-hub has sprung up to route around this fault, and provide people with the resource.

Practically, there are real people making decisions who are involved in this process, but on an abstract level it feels like this is somehow inevitable? It feels like an emergent behavior from any sufficiently advanced network, and any attempt to stop it seems futile.

I guess another important part is that it's the universities' connection that scihub 'uploaders' tunnel through that are key - they have the special network access to DOI resources that make it free for on-site university users to get anything for free, and then the 'uploaders' relay them to outsiders. That's a difference between papers and something like 'scihub for any pdf ebook via ISBN' - there's no special network that can get all ebooks for free just via ISBN - except maybe the Library of Congress ;)

> I guess another important part is that it's the universities' connection that scihub 'uploaders' tunnel through that are key

Does this happen in realtime? I read that sci-hub sometimes uses credentials from legitimate users, but in that case they would be easily identified.

So, perhaps someone with access submits papers on request? And if so, what about watermarking?

> And if so, what about watermarking?

years ago i recommended to scihub that they should use https://github.com/kanzure/pdfparanoia but they weren't too concerned.

Plenty of the papers I've gotten from scihub are watermarked. The "downloaded by [IP] at [timestamp]" footer could probably be stripped, but they don't appear to do so, or at least not consistently.

A lot of papers get ingested into sci-hub while they're still in the preprint/early-access stage. Sci-hub doesn't appear to reingest papers after the final version is published. It's one of my very few gripes about the system. A more quality-focused system would probably handle this edge case.

This seems more like a versioning issue. Can't papers that pass review get a new DOI? I'm thinking something like semver for scientific papers or scientific data, more generally. DOIs even look like semver versions, so why not add more space for the different stages of publishing?

I'm talking about papers like this:

http://sci-hub.cc/10.1021/acs.jpcc.7b06335

It's already passed peer review, but the final version will be tidied up with additional editing and will have e.g. the line-by-line numbering stripped out. I see sci-hub versions of papers from months ago, or older, that never update to the final edition. The content doesn't differ much but the final version may have better-placed figures, corrected typos, and other minor improvements.

If different versions of a paper have different DOI then the DOI is broken and should be replaced by something better.

I consider it a heresy that some sites (fuck ResearchGate, Academia.edu) consider it fine to alter the PDFs. Document integrity should be the holy grail. Files should be final and never get touched. If they are, they are different.

Wait, aren't those two statements opposed to each other?

Different versions of a paper should have different DOIs. A DOI is fundamentally quite abstract, and there is no reason that an object-appropriate versioning scheme can't be laid on top. Figshare, for example, uses a versioning system where there is a single DOI referring to the overall artefact (e.g. 10.6084/m9.figshare.2066037) and a versioned-by-suffix DOI referring to a specific version (e.g. 10.6084/m9.figshare.2066037.v16). That works really well IMO.

ArXiv itself does this. See the 'cite as' section of this one: https://arxiv.org/abs/1707.08475, which has the note

arXiv:1707.08475 [stat.ML] (or arXiv:1707.08475v1 [stat.ML] for this version)

Oops, yes of course. That's what I meant. Typo. :)

The publisher is using DOIs incorrectly.

We all miss oink.

sci-hub is not like it, but it's more curated that tbp. You can enter a paper's URL or DOI and it'll take you right to it.

I think there is curation that should be build on top of it, especially if the concept of "review" could be moved to an online system.