Hacker News Reader

Comment by homarp | Hacker News Reader

The article does not discuss sci-hub unfortunately.

list of the 176 vanished is here: https://github.com/njahn82/vanished_journals/blob/master/Dis...

What would be really useful is to know the average citation count of these journals.

As someone who hates to see this stuff disappear, there is still a cynical person inside me that knows there are a glut of journals that are often used to bump publishing count for professors trying to get tenure.

That cynic inside of me also realizes that a subset of the journal business is a bit of a scam anyway since frequently authors have to pay the journals to include their paper and then the journals charge an exorbitant rate to get access. Have you tried to buy a journal article lately ($35 for one paper!) -- yeah, neither have I.

true and true. "Sci journals" are everything but science

Not to mention many of these OA journals were only created to scam researchers. They’d create a fake front and pretend they’re a prominent journal. Can’t recall them specifically but there were a few news clips about them a few years back.

Can people not try to make a scam out of everything??

Problem is that almost anything that can be exploited eventually will be.

In this case the front may be specious instead of an absurdity but it’s no less a front.

1. https://en.wikipedia.org/wiki/Predatory_publishing

2. http://web.archive.org/web/20170128194124/http://www.chronic...

One can be a tiny bit more charitable and say many of them provide a nearly worthless service (slapping your manuscript on their fly-by-night website) while misrepresenting themselves as providing a much more valuable service (peer review, dissemination to an audience that actually reads their journal, etc) and charging accordingly (thousands of dollars per manuscript).

Perhaps it's arguable whether that is to be called a scam, but certainly it is a rip-off.

It's worth noting that citation counts are an increasingly poor metric of paper quality (and always has been).

There are multiple works to show that the rise of search engine like Google Scholar have meant that researchers are increasingly citing the same papers, because their searches are all returning the same thing.

Meanwhile, there are some "sleeper" papers that are super relevant to a lot of works, offer great insight, but by virtue of their low search ranking, never get cited.

That's not to say there isn't a fair amount of unremarkable research. It's just that it doesn't always correlate with citation count.

To possibly find “sleeper” papers check out the “show similar” feature of this arXiv mirror on condensed matter physics. https://cond-mat.abbrivia.com Search for some keywords and then surf by “show similar” on relevant articles.

Thank you, now I know the term to associate with one of our papers written by one of the students, a "sleeper paper". The twist is that Google ranks it consistently high in the very first page in Google search (but not Google Scholar) for well over a year now for two common keywords. Somehow it never get a good citation count mainly a few handful self-citations even it has over 99% classification in all the important metrics namely accuracy, sensitivity and specificity.

Curiously I have another co-author paper by another student that was published around the same time (same topic different application) and as of today it has over 60 citations count. It even cited by a Stanford University researcher but only with classification accuracy of below 90%!

I think this is a classic case of a sleeper paper where you have excellent results but then nobody want to cite you because it makes their paper look bad or negate their (fake) claim of novelty. FYI, most of the papers in the field use the same venerable open online database that make it easy to compare your results against others.

Even the most worthless articles and the worst pay-to-publish scam journals should be preserved as historical documentation and negative references.

Link seems to be broken: fixed version may be https://github.com/njahn82/vanished_journals/blob/master/dat...

Yes, correct. We cleaned up the repo and removed outdated files that are not needed to replicate the figures and tables to avoid confusion. The dataset can also be found here: https://zenodo.org/record/4014076#.X1lgTbexVkw

A list of 176 but it's an excel file, why?

Dataset is published on Zenodo (as .csv). https://zenodo.org/record/4014076#.X1kj-rexVkw

I'm getting a 404 at that link, now.

Reply on news.ycombinator.com