What I find the most fascinating is that a single person, possibly with the help of a few select others, is able to curate, index, host, and serve 85% of scholarly literature[0] which otherwise would be inaccessible to many.
80 million papers are available on Sci-Hub to its half a million daily users.[1] That is around 80 TB of data, and 2 TB of daily outbound traffic.[2]
Given the fact that Sci-Hub has achieved all those whilst being forced to operate under the radar, I have massive respect for Elbakyan first and then all the others that have helped it survive. Not only Sci-Hub is an undeniable disruption, but also an amazing technical feat that many startups would envy.
Something that came up often when talking to people at my college was that our extremely expensive contracts with publishers/journals were to support the complex infrastructure required to maintain a service like that. We're talking multi million dollars per year per university per journal contracts.
Reviewers are volunteers. Hosting costs, as shown by Sci-Hub, are negligible. How much does it cost to replicate a file system across 5 to 10 data centers, have a 10G link to each data center, and run a few nodes to serve web traffic?
Probably a small fraction of the budgets we pay for journals.
> P.S. Seed its torrents!
I know no one will give legal advice here but: is this legal to do?
"A reminder that Elsevier made
$6 BILLION selling your academic journals and articles behind paywalls, and made more profit than Amazon, Google and Apple every year for YEARS...
And paid the academics who wrote the articles $0
And paid the reviewers of the articles $0"
> our extremely expensive contracts with publishers/journals were to support the complex infrastructure required to maintain a service like that
> Hosting costs, as shown by Sci-Hub, are negligible
Yes. The truth is these problems and costs are created by the copyright industry. Since copying is illegal, they have set themselves up as the only legal source for this material and that requires a lot of storage and bandwidth.
All they have to do is place their copyrighted works in the public domain. The storage and distribution problems will solve themselves immediately and at zero cost to them. People will literally do their work for them.
> Probably a small fraction of the budgets we pay for journals.
Well motivated enthusiasts in their free time can achieve 100x what middle-grade 'digital consultants' can do in the same time...
Building sci-hub in a corporate world, even leaving out the legal issues, would be very expensive. Begin by thinking how many 30 person meetings you're going to need to have to decide how to arrange the tender process...
In Germany you'll get sued very quickly for offering others' copyrighted content, there's a whole industry there of lawyers mailing you that you've violated their clients' (copyright holders) rights but if you pay the around 1000 Euroes "damages" and sign a letter agreeing to not violate their rights again it'll go away. If they catch you seeding/uploading a part of another of their copyrighted content, then they'll ask for even more money.
So if you think seeding is the right thing, you'll probably end up really broke and unable to pay your bills there.
The other side of this are lawyers profiting off people who get these letters and charging money for basically copy-pasted emails on how to respond.
That's true. Putting tax funded research studies behind corporate paywalls also is in many contexts and jurisdictions. But it's funny how that one doesn't ever get prosecuted.
It also shed some light on the inefficiency / inefficacy of most production groups .. a few persons with a real motive ended up doing more than many large companies.
Nothing new, it's obvious to many how much drag and cruft lies in the majority of businesses in the world.
I work in corporate software development, and the incredible inefficiencies astound me every day.
There's a lot of useless work mandated by law as well... I've spent years of my life developing useless anti-money laundering software because a lawmaker somewhere thought of a convoluted potential possible money laundering insurance fraud scheme (while ignoring the actually used ones).
The insurance company had upwards of TWO THOUSAND developers working... and the software was incredibly bloated and slow crap. It's mind boggling.
I have a NAS which I could probably open up and seed a lot of Scihub content from. But what are the odds that something bad would happen to me, and how bad? Like, a fine? or a raid and jail time?
if you are in a developing nation, which most of users likely are, then seeding or downloading illegal torrents won't be any problem. But I don't know about others.
80 million papers are available on Sci-Hub to its half a million daily users.[1] That is around 80 TB of data, and 2 TB of daily outbound traffic.[2]
Given the fact that Sci-Hub has achieved all those whilst being forced to operate under the radar, I have massive respect for Elbakyan first and then all the others that have helped it survive. Not only Sci-Hub is an undeniable disruption, but also an amazing technical feat that many startups would envy.
[0]: https://greenelab.github.io/scihub-manuscript/
[1]: https://translate.google.com/translate?sl=auto&tl=en&u=https...
[2]: 3 articles × each 2 MB × per 400k users ~ 2.4 TB.
P.S. Seed its torrents!