The lock free / wait free datastructure literature is not an especially fertile land to go searching in. The most common flaw is probably assuming you're runnin

The lock free / wait free datastructure literature is not an especially fertile land to go searching in. The most common flaw is probably assuming you're running on a garbage collector with better properties than the structure you're describing. A close second is assuming CAS on two unrelated addresses exists. A fair few are just wrong as these things are difficult to design by thinking hard and people don't tend to write the formal proof of correctness. Even those that look right, actually wrote the algorithm in something approximating a formal language and don't assume they're built on magic, are difficult to search for and to compare with other designs. At this point I'm just going to continue DIYing the things and previous inventors can only have credit if I know their thing exists and can reasonably determine whether it's the same as my construct. Which will presumably annoy people much like the authors of this post. I note that I hadn't seen the left/right paper prior to this thread so have no idea whether I've recreated it without proper attribution. Seems likely that'll be the case for other people writing stuff too.