800 GiB torrents with 1500k public domain paywalled papers from before 1923 : DataHoarder

"Happy public domain day to all! A special thought for people in the USA, who until yesterday suffered a 20-years-long winter of ever-expanding copyright.

Today we return to the racket of academic publishers who abuse copyright to enslave thousands of researchers and leech billions in public funding from cultural institutions every year. Following my release of public domain papers and seeing the format of some other releases by other users, today I bring your attention to Scholarly works published until 1909(torrent) and in 1909–1922 (torrent).

Please download and seed the torrents above! Or if you prefer you can add the hashes/magnet links directly, but not all clients support the web seeds provided this way....

These datasets contain 1,518,078 PDF files from various sources, often with added OCR. They were all published before 1923 in international journals. I'm not providing legal advice, but if you consider them simultaneously published to USA they should all be in the public domain in the USA. Yet, publishers apply indiscriminate copyright statements to the contrary, which may constitute copyfraud, and lock nearly all of them behind paywalls or other hurdles, hoping to milk some more profit for who knows how many centuries.

You can also download PDFs by individual publisher, going by their DOI prefix and checking the full list of DOIs (1909, 1923). Internet Archive can also support the direct download of individual files inside the ZIP files but that's probably best handled by other repositories...."