Hacker News Reader

Comment by d33 | Hacker News Reader

One thing that worries me with Sci-Hub is that even though they put all of their PDFs on the torrent, it's too much to mirror by one individual. It would be nice if they implemented some scoring system that would help others mirror the most popular papers or if they worked like thepiratebay - just associating names to magnet links. It would be much easier to help work around the single point of failure then. Right now those huge torrents have as few as 4 seeders.

View on HN

It's not the popular papers that are important. Those will be well spread already. It's the unpopular "rare" ones that need help.

You mean the libgen torrents or is there something spefically branded sci-hub?

It's possible to solve, you just make users download a zip file which includes 10 papers including the one they're interested.

Bitorrent kind of does this anyway, because in order to download one file, you need to download the entire block it's in. In a case like this where you have a large torrent with many small files, that will actually cover quite a few other files. If you ordered them strategically, you could even try to make sure every block has a fairly popular file in it.

Yeah, good idea. Sort the papers by popularity, and when grouping them find the distribution that minimizes the distance between any paper and the top-popular ones for atomic block size.

I don't think you need to do anything that complex. Just find the N most popular papers and make sure there's one in each block.

Or thinking about, you could put them at block boundaries so people grab two blocks when they download them, in which case you'd only need half as many popular ones.

I don't think that would accomplish much. People would just delete the 9 papers that they don't care about rather than maintain them as clutter. And even if everyone did keep the extra files, would use would they be? It would be impossibly difficult to reconstruct the archive from them.

It would probably be better to break down the torrents into smaller, more coherent, and organized collections that contain all associated metadata (e.g. a collection of biology papers from a certain time-span or set of publications). That way you might incentivize people to maintain smaller personal-use archives that could be recombined back into the whole, if the need arises.

You could perhaps even build a system that could automatically generate, publish, and consume torrents with a particular (metadata + content) layout, but that's not restricted to a particular set of torrents. Then it could just consume certain topics from a feed of new torrents, from various sources, on the Pirate Bay.

Why, though? It's a waste of bandwidth to force someone to accept a bunch of information they don't need and didn't ask for.

I've always just thought we'd make a system where storing some extra information was just the price of entry. That is, in order to torrent the papers you want, you need to first share an extra gigabyte or so of papers chosen via some algorithm to ensure their availability.

I'd feel pity if you wasn't trying to get something for free...

Heaven forbid you help propagate lesser used stuff as payment for getting access to the few papers that you feel entitled to get for free...

> Heaven forbid you help propagate lesser used stuff as payment for getting access to the few papers that you feel entitled to get for free...

I'd more than likely just delete the unusable stuff after setting aside what I actually needed.

So, yeah, waste of bandwidth to send in the first place.

The idea, especially with Torrents, is that you are sharing what you downloaded...

So even if you download 10 files, and then delete 8... BEFORE You delete those 8, you can "share" them for a small period (or until you have positive "Download to Seed" ratio), thus helping others.

Pay it forward and what not...

That's the theory... lots of people download and stop thus leech without helping.

I believe that when you download only a single file from a large torrent , you still help the torrent with your file.

So if they just put torrent links in their search engines, and ask for people to install a browser torrent extention like [1], it could work pretty well.

[1]for firefox: https://addons.mozilla.org/en-US/firefox/addon/torrent-torna...

https://ipfs.io/

Yeah, the thing is that SciHub controls the information now and it's up to them how they secure against failure they are. All in all, if we were to adopt IPFS, who would download all the terabytes of papers? Who and how would index them?

If they used ipfs as backend, everyone could sync and serve whatever parts they deem important. At least if I understand it correctly. ipfs is way too complicated for me.

I would much prefer an anonymous network anyways.

Freenet offers an anonymous, censorship-resistant distributed data store system like IPFS. It is very well suited to something like the Sci-Hub archive.

This is a work in progress.

> Right now those huge torrents have as few as 4 seeders

Any ideas on fixing this? I've tried offering her storage, bandwidth, or physical delivery of hard drives, but so far no bites (specifically, outright refusal).

Install a torrent client, download the torrents, keep seeding?

Is it safe to do? I mean on the legal side.

Safe? No, it's probably copyright infringement. But you have to consider that so is reading them on sci-hub.

The question is whether you want to engage in this particular act of civil disobedience and bear the risks or not.

The risks can also be reduced by living in the right jurisdiction or obscuring your identity, e.g. by renting a torrent seedbox with bitcoins.

No.

What's wrong with setting up a mirror at archive.org? After all, they /are/ a library, so they get some immunity from copyright. Or can we not do that?

> What's wrong with setting up a mirror at archive.org? After all, they /are/ a library, so they get some immunity from copyright. Or can we not do that?

Pretty sure no-one gets "immunity from copyright," at list not to the point where they could mirror Sci-hub with impunity (at least no on in the US). I'd very surprised if archive.org did not respond to DMCA take-down requests, etc.

On the other hand, I believe I've read before that DMCA takedowns only cause content to become inaccessible publicly but not actually deleted, which in the case of SciHub sort of defeats the purpose; however, it does mean archive.org retains an "archive" of the content, presumably until its copyright expires or copyright laws change.

> I've read before that DMCA takedowns only cause content to become inaccessible publicly but not actually deleted

This is a violation of the takedown request, and can result in harsher penalties for the host. It's not a great example, but MegaUpload was faulted for doing just that: Removing the link between a url and the content that url pointed to. Granted they did deduplication so there could be other links from other urls pointed to the same "illegal" content, which is what the USA went after them for, but the fact remains that it's a pretty clear grey area whether or not the host can retain that data after someone claims rights to have it removed for their own future benefit.

Looking at the text of the law, it always uses "remove, or disable access to" or similar language: https://www.law.cornell.edu/uscode/text/17/512

> presumably until its copyright expires or copyright laws change.

Then what's the point? The whole point of Sci-hub is to distribute copyrighted papers, "archive" would be inaccessible until it is useless.

The point would be to keep an archive of scientific papers, but prevent access to them in case SciHub goes down.

Do you have a link for the torrents? I couldn't see anything on the main sci-hub site.

My suggestion for such filesharing is sending her money to anonymously send some people hard drives or access to file upload sites that have a copy. They can make their own copies that spread. And so on.

Reply on news.ycombinator.com