> Is there an actual project calling itself that, or is that just a figure of speech that the writer of the article used (inaccurately)? Because abandonware, unpublished or reverse engineered software is not automatically open source, which has a very specific legal definition.

Open source does not have a "very specific legal definition". Many different people use that term to mean many different things, the chief commonality tending to be "can you read the source code". Yes, OSI has a specific definition that goes beyond that, but OSI is not a legislative or judicial body, nor do they have authority from any such body to be involved in regulatory activities (contrasted with, say, the Amateur Relay Radio League, or Civilian Marksmanship Program). And personally I hew more towards the FSF's views on the matter than OSI's.

Anyway, it's pretty common in the open source world for a specific name to be trademarked, or assets to be copyrighted, while still making the source itself open. Some examples include SimCity vs Micropolis, Chrome vs Chromium, and Firefox vs IceCat. I'd argue that such projects still count as open source, because the source itself is open, even if you can't use the name or non-code assets. Non-code assets like names, graphics, audio, and documentation are commonly available under different terms or licenses anyway, e.g.: https://wiki.wesnoth.org/Wesnoth:Copyrights Pretty much everyone (OSI, FSF, etc.) agree that code-specific licenses like GPL, MIT, BSD, and Apache don't make so much sense when talking about things that aren't software, hence the existence of things like GFDL, CC, etc.