Borg said:
TIL that IBM patented a technique created by Euler.
In the late 1980s, working for a university, I wrote some terminal graphics routines for the IBM PC which among other things drew arcs without needing multiplication or division, and I was told I couldn't include them in the terminal program I was selling because the algorithms had been successfully patented by someone else, despite the basic stated principle that algorithms are not patentable. That was a shock and very annoying; I ended up omitting that functionality from the program, making it emulate a less capable device (which still worked, but required the mainframe programs to draw the arcs as sequences of line segments, impacting performance).
When I worked for IBM later, I was told that even though pure software and algorithms should not be patentable, others had patented such ideas by claiming the patent applied to an example hardware implementation, so we needed to protect our software by the same approach.
A few of my IBM software inventions (mostly relying on particular hardware features) were patented (including at least two which I thought were obvious but which were initiated by others who included me as a shared inventor). I also participated in invention reviews to decide whether other ideas should be submitted for patents. (One of my reviews consisted solely of "That's the DOS APPEND command").
To start with (early 1990s) we were really careful with our own searches for prior art, and only let through ideas which really seemed novel, non-obvious and useful. (With my diverse range of computer experience, I tended to find prior art rather too easily, which was discouraging for the applicants, as was my general attitude to trying to patent something which appeared to be pure software, so I was quite relieved when I stopped being asked to do reviews). The review process took a lot of internal skill and effort, and others pointed out that it's really the job of the patent examiner to check whether it's valid, so we reduced our internal checks and started submitting more stuff that hadn't been fully checked. It then turned out that patent examiners were so overloaded that they started to let stuff through without much checking, so that in the end real checking would only occur if someone decided to challenge the patent, and that's probably still the situation now. (I was still sometimes asked by IBM whether I was aware of any prior art which could help to challenge someone else's dodgy patent).
So the position at IBM when I was there appeared to be that if you do anything you consider ingenious, you're expected to submit some patents to protect it, preventing anyone else from doing so. The pre-submission checks are no longer very thorough, and neither is the external patent review process, but if there is a problem the patent can be overturned if you have the resources to challenge it, which is somewhat unfair as that is usually a costly process.
In 2001, an Australian managed to patent the wheel, although it was revoked in 2014 after a challenge!
