In my previous post , catalyzed by Sci-Hub, https://blogs.ch.cam.ac.uk/pmr/2016/04/30/a-commentary-on-sci-hub-1-scholarly-publishing-is-broken/ I argued that scholarly publishing is completely broken. It’s now lost a huge amount of respect, it’s unwieldly, unfair and mired in bickering. It pays no attention to readers. It’s becoming a write-only system where authors write not to communicate but for glory – self advancement. There’s no clear political goal …
… and no clear technological goal.
And that’s the problem.
Because we desperately need the ability to search and analyze the scientific and medical literature in a 21stC manner. While we’ve been creating our http://contentmine.org we’ve discovered many researchers who have to “read” 10,000 papers in a day or two. They use 20thC methods – click and read – taking weeks where they should take hours. ContentMine software (completely Open) has been built to solve this problem by filtering out the papers you don’t want – often 90% of the first search. (and it does much more – it can extract complex objects). It’s Open to everyone and it works (see previous posts).
When I came to Cambridge I had the vision of building an “artificially intelligent chemical reader” part of which was the World_Wide_Molecular_Matrix a system for capturing and sharing versioned semantic chemistry. Bits of it are being built in ContentMine . I built systems where I could draw chemical formulae by speaking to the machine. We’ve built the de facto tool for chemical name recognition (OSCAR) and interpretation (OPSIN). I thought it would take 5 years to create my chemical amanuensis – scholarly assistant. With help from the publishers and scientists it probably would have. Now, after 15 years, it’s still a dream, frustrated by stagnant thinking on all sides, and deliberate opposition (e.g. nullifying European legislation).
So Stackoverflow, Github, Bitbucket, Apache, GNU, Jenkins, OuterCurve, Mozilla and many others are creating the human-machine technology of tomorrow. This encourages innovation from predictable and unpredictable sources. It works – it’s exciting and we are all part of it.
In contrast the Scholarly publishing industry has created nothing in the last 20 years. (The Scholarly Kitchen hailed the “big deal” (a pricing strategy to increase sales) as one of the greatest achievements of schol pub).
20 billion dollars per year – that’s 200 billion since I started at Cambridge – and nothing positive to show for it.
The current technology of the mainstream publishing industry is just awful. Really awful. It’s often built by outsourcing parts to people and companies who do not care how the result is used. The methods used – awful PDF and really awful HTML – are for the publisher’s convenience , not for the reader. And every publishers complains about how awful the tools are. They can’t change, they can’t innovate, they’re locked in. Add that every publisher feels they have to use a different technology to differentiate themselves from the others and it’s a complete tower of Babel. (I have spent 2 years of my life trying to solve this awful mess – and ContentMine can untangle a good deal.)
What’s even worse is that most of the publishers spend effort on STOPPING people reading the literature. The obstacles to getting to a paper grow every month. These include (from my own experience):
- Deliberately bad PDFs.
- Pixel maps rather than characters.
- “Glass screens” that can’t be copied (Readcube from Nature/Springer).
- Captchas to stop readers after 25 papers (Wiley, i.e. 400 Captchas for a literature review).
- Monitoring every download and requiring libraries to stop researchers. (Elsevier, Wiley).
- Automatically cutting off 200 universities for a single click (Amer. Chem Society).
Why does this matter?
Because there is so much we are missing out on. New medicinal knowledge, new ecology, new astronomy, materials, chemical reactions, … and innovation…
I should be able to ask a computer (in speech):
“Find me all chemical compounds that occur in Lantana species south of the Wallace line and compare their chemical and plant evolution. What types of compound might we see in the future, particularly due to invasive species?”
And get a result in minutes… it’s not as hard as it looks. It’s knowledge-driven science.
(Sadly All I WILL get in minutes is a cease-and-desist letter from publishers demanding that I shouldn’t “steal their content”.)
So because we cannot innovate in this area we are 20 years behind the mainstream.
So why do I want Sci-hub? (Note carefully that I haven’t said what I am going to do and, until I do, you cannot judge my intention. I haven’t said I’m going to use it. You’ll have to wait till the next blog post).
I want Sci-hub because it’s technically BETTER than anything else we have. Much better.
And it’s the perfect complement to ContentMine.
Sci-hub has all the world’s scientific knowledge in one logical place. It doesn’t matter that it’s spread over Torrents and other fragmentation – logically it’s all there. And it’s run by someone who knows what she’s doing technically – unlike many publisher sites. And, I assume, she and colleagues will be receptive to technical requests and suggestions. (No one has any chance of getting conventional publishes to innovate).
Using Sci-hub would advance my and ContentMine technology enormously. ContentMine and Sci-hub fit together perfectly – because they are both designed with the 21stC mentality. Because they react to what readers want. Yes, READERS; the marginalised community of scholarly publishing. 21stC projects create a community round them. They are organic and vibrant. They respect machines and humans equally.
ContentMine + Sci-hub could be the greatest search engine in scholarship, especially for science, technology and medicine. Because it’s semantic. Because all the literature is trivially accessible in one place and one format. I don’t know of anything that remotely comes close. We can search and index diagrams – extract 15 million chemical reactions a year. (Even if a publisher tried to develop it they could only use it on “their own” content.)
BUT! BUT! BUT!
But for many, including the law, Sci-hub is forbidden fruit. Run by She-who-must-not-be-named. The arch-pirate. The criminal. (These terms are used). Peter Murray-Rust cannot use it (and I haven’t). ContentMine cannot mine it (and we won’t). We’ve looked at the legal and political aspects and I’ll analyse these in a subsequent post.
But 21stCCitizens – me, ContentMine, taxi-drivers really really want Sci-Hub.
The only things stopping us are copyright law, prosecutors and an intransigent, uncaring, out-of-touch, money-driven and self-seeking publisher-academic complex.
I’ll deal with the politico-legal in the next post.