Is it just me or is passing off things that aren’t FOSS as FOSS a much bigger thing lately than it was previously.
Don’t get me wrong. I remember Microsoft’s “shared source” thing from back in the day. So I know it’s not a new thing per se. But it still seems like it’s suddenly a bigger problem than it was previously.
LLaMa, the large language model, is billed by Meta as “Open Source”, but isn’t.
I just learned today about “Grayjay,” a video streaming service client app created by Louis Rossmann. Various aticles out there are billing it as “Open Source” or “FOSS”. It’s not. Grayjay’s license doesn’t allow commercial redistribution or derivative works. Its source code is available to the general public, but that’s far from sufficient to qualify as “Open Source.” (That article even claims “GrayJay is an open-source app, which means that users are free to alter it to meet their specific needs,” but Grayjay’s license grants no license to create modified versions at all.) FUTO, the parent project of Grayjay pledges on its site that “All FUTO-funded projects are expected to be open-source or develop a plan to eventually become so.” I hope that means that they’ll be making Grayjay properly Open Source at some point. (Maybe once it’s sufficiently mature/tested?) But I worry that they’re just conflating “source available” and “Open Source.”
I’ve also seen some sentiment around that “whatever, doesn’t matter if it doesn’t match the OSI’s definition of Open Source. Source available is just as good and OSI doesn’t get a monopoly on the term ‘Open Source’ anyway and you’re being pedantic for refusing to use the term ‘Open Source’ for this program that won’t let you use it commercially or make modifications.”
It just makes me nervous. I don’t want to see these terms muddied. If that ultimately happens and these terms end up not really being meaningful/helpful, maybe the next best thing is to only speak in terms of concrete license names. We all know the GPL, MIT, BSD, Apache, Mozilla, etc kind of licenses are unambiguously FOSS licenses in the strictest sense of the term. If a piece of software is under something that doesn’t have a specific name, then the best we’d be able to do is just read it and see if it matches the OSI definition or Free Software definition.
Until then, I guess I’ll keep doing my best to tell folks when something’s called FOSS that isn’t FOSS. I’m not sure what else to do about this issue, really.
In the case of Machine learning the term has sort of been morphed to mean “open weights” as well as open inference/training code. To me the OSI is just being elitist and gate keeping the term.