News for hackers. Stuff that matters.
Re:Sci-Hub statistics and database by aurizon on Saturday February 12, 2022 @06:47PM (#30317568)
The publishers now encode their papers with individual identifiers, that generationP calls UUIDs on all pdf's. That means they can trace it back to the institution(and perhaps the actual prof?) They have then sent nastygrams to threaten them with fees or loss of access to the institution - potentially serious punishment. What is needed is a way to run the papers through an OCR recognition program to create renewed text and to further process to randomly vary a large number of adjacent letter spacings. (this is called 'micro kerning' and it allowed a second order of unique document recognition where the document is scanned by the journal to look for these fractional kerning gaps in the letter spacing which can lead back to the institution). I suppose a program flow could be made to OCR + random micro kerning changes - it would take time, but once set up it would be a rapid computer based document flow process. Photos/charts could be sanitised as well, but legends etc on them would also need OCR and micro kerning adjustments. With 25 words of text, each with 6 letters, micro kerning can easily create 10,000 unique ID's, easily enough to cope with all subscribed institutions - and that is on one chart/photo. I suspect the subscribing institutions have acted with firm words to their profs and grad students to block this. This can easily be told to us if a few people in assorted institutions let us know if they have been read firm words about this?

I am not sure how Sci-hub can get past this, unless they get a good Indian court ruling and can use Indian friends to scan printed copies of journals - if they exist in this online age?

Re:Sci-Hub statistics and database by hda2 on Saturday February 12, 2022 @11:17PM (#30319166)
> I am not sure how Sci-hub can get past this

Quite trivially, actually, thanks to the good old analogue hole: https://news.ycombinator.com/item?id=30084193

All sci-hub would have to do in this case is download the same paper through three or more accounts (different institutions, networks, countries?) at three different times, rasterize them and keep the common denominator. If a pixel has no common denominator, they'd have to fall back to a default value. This is by no means a perfect method and it has its weaknesses and pitfalls, but the resulting PDF will be far less useful as a means to de-anonymize accounts using information from the PDF itself.

Publishers still have other sources of information to de-anonymize accounts if the multiple accounts/downloads aren't truly isolated from one another.

Re:Sci-Hub statistics and database by aurizon on Sunday February 13, 2022 @07:09AM (#30321182)
Yes, that would mush their secret sauce...
Re:Sci-Hub statistics and database by lucb1e on Saturday February 12, 2022 @09:49PM (#30318664)
Makes me wonder how much they'd have to offer me to accept the task of implementing such tracking algorithms to modify other people's scientific papers for the purpose. Certainly I won't be the cheapest, but still. Either some developer is very vested in the idea of keeping science a secret or someone got a very nice bonus.

Edit: also the sysadmin that keeps this database safe without 'accidental' data loss on UUID to downloader mappings.

Re:Sci-Hub statistics and database by aurizon on Sunday February 13, 2022 @08:39AM (#30321686)
It seems to me that this type of tracking, under the guise of guarding rights is against the GDPR = unlawful tracking that they are harassing Google and FB with? Is not justice's sword 2 edged?
Re:Sci-Hub statistics and database by lucb1e on Sunday February 13, 2022 @02:43PM (#30324899)
(Note that this is all written from a European perspective, applies European Economic Area law and human rights as defined by the Council of Europe (which includes Russia, to everyone's surprise). I know that in the USA everyone is much more pro frontier justice, for example when it comes to pervasive and continual monitoring of employees while at work.)

I think it's very arguable that they have a legitimate interest here. Privacy has always been a weighing of interests, at least that's how I've always heard it explained by the Netherlands' face of digital law (Arnoud Engelfriet) also back in the days of WBP (the law from ~1995 that is 97% the same thing as GDPR), also in light of the European Convention of Human Rights (article 8 is a right to privacy).

A common example is filming the road: illegal, but if you park your car in front of your house and there have been car fires in your neighborhood lately, then it can be justified.

Filming employees inside a warehouse: invasion of privacy (illegal) but if there have recently been thefts from a certain part of the building then it's justified to hang up a camera there, introduce a lock that registers who went there at what time, or some such. (With adequate security measures so only authorized people can use it for the intended purpose.)

Personal example: monitoring everything I do on the company network is illegal, but because I work in a business where secrecy is important (security consultancy) it was considered justified to do spot checks, tell every employee upon entering into the employment contract that spot checks are a thing, and inform the subjects of spot checks after they were part of one. Transparent but still effective.

The two things to consider (iirc) are:

- Do my rights weigh heavier than the other party's right to privacy? (e.g. car fire is a fairly big impact on your right to the peaceful enjoyment of his possessions)

- Is there any other way in which I could achieve this goal with a lesser impact on the right to privacy?

In the case of Elsevier, from what I heard this whole scheme is a big mafia-like practice (wouldn't want to be published in a niche corner nobody reads now would you?) and so in my opinion it's entirely unethical to support (work for) them in the first place, at least in any role except one where you think you might be able to nudge things in the right direction. But I could see how a judge says: well, that's how today's law works, that you have moral objections is something you can take to your favorite religious leader and lament about, not a court of law.

If I'm being fair, there isn't even really an invasion of privacy because PDFs don't have executable code (usually) that can track you. Rather, they need to hide it somewhere so that, if it appears on the pirate bay, they can read out the ID and see who the perpetrator is. More like a criminal investigation using a fingerprint on a glass, and less like a cookie actively sent with every action you perform on a website.

TL;DR: GDPR applies, but it probably doesn't make this database illegal. It's not a loophole by which a person can say no to literally everything. (Would be cool if you could require the police to stop using your fingerprint in a legitimate investigation.)

Still, if I were that sysadmin... I probably wouldn't 'drop table elsevier', but I'd rather live off government benefits than support that scheme.

Re:Sci-Hub statistics and database by aurizon on Monday February 14, 2022 @08:26AM (#30331710)
Thank you for this cogent analysis. It sounds valid as they probably notify their subscribers that they forbid document release to the wild and use technical means to measure it. The only hope is the usual request to the authors, who have traditionally sent papers to all requests, of course the profs can also pay 'blotgelt', and as long as the fees are small enough or the prof can afford, file it as an open paper. I am encouraged that the open paper concept is gaining traction - in the days when Nobel was alive, his Nobel Prize was in fact created on this basis of open papers freely sent to others - sadly, had he known, he may well have included the open paper concept into his legacy to prevent the blatant rent seeking empires we now see. That said, curation and acceptance/publication of papers is a service of value and it is needed, at a lower cost as the journals do a job of work to keep totally crappy AI/idiot created papers that get into the totally open journals who are unable to deal with the blizzard of papers they face - the bad get through. There are also the fake fee based online journals where a distressingly small fee gets you online for citation by your coterie of anti vaxxers, nutrition gurus, etc etc. - essentially insoluble, save by intelligent readers who often shed this burden to avoid the waste of time. Fortunately the 'contents **' publications winnow most of these out.
Re:Sci-Hub statistics and database by robwwilliams on Saturday February 12, 2022 @07:13PM (#30317729)
No “firm words” that I heard in my neck of the USA woods.

I was struck by very low number of German downloads. Did I miss something?

Re:Sci-Hub statistics and database by aurizon on Saturday February 12, 2022 @07:45PM (#30317926)
No idea about Germany, but they may well be more law abiding?
Re:Sci-Hub statistics and database by medstrom on Saturday February 12, 2022 @09:13PM (#30318467)
No, that's absurd. More likely a different factor. Some googling and lo and behold: starting with Germany, the universities of several European countries cancelled their agreements with Elsevier, pressuring Elsevier to give them better deals, including open access by default. I imagine it lessens the need for Sci-Hub.
Re:Sci-Hub statistics and database by aurizon on Sunday February 13, 2022 @07:21AM (#30321229)
A good beginning
Re:Sci-Hub statistics and database by pishpash on Sunday February 13, 2022 @02:08AM (#30319892)
Do a reflow of the entire document? As long as it's semantically recognizable it can be done. Figures are harder. ML can really help here: someone make a pdf2tex!
Re:Sci-Hub statistics and database by trasz on Monday February 14, 2022 @09:05AM (#30332109)
Does this mean it is trivially easy for any student or employee to make their institution stop paying publishers by getting it banned?
Re:Sci-Hub statistics and database by aurizon on Monday February 14, 2022 @12:23PM (#30334762)
LOL, student sabotage!. I think the paywalls would extract a penance that maintains their cash flow - remember, Greed is their mantra, may it never ever cease.