We also don't know what the movement would look like today if it weren't dominated by their personalities (and their personality conflicts) in its formative days. It's pretty clear that a lot of people were interested in the ideals behind their movements and contributed in massive ways to the success of the movements - it's not clear that either of them were needed to get the movement off the ground (see also, Great Man Theory).
I seem to recall a post on his blog where he addressed this point. IIRC, his conclusion was that if it had not been him, then then the culture/movement/whatever would have eventually invented someone else like him. That is, someone would have eventually noticed the same things, come to the same or similar conclusions, recognized (or at least thought) they had the social skills/aptitude needed to successfully promote the ideas to the wider world, and gotten started. I can't disagree with that. The question to me is whether that would have happened around the same time, or if it would have taken several more years for the same things to click for someone else.
What it would look like: an army of docile microserfs beavering away for free to make FOSS work better for Google, Amazon, Microsoft, Facebook, Oracle, IBM, ... and the odd government who wants a crypto backdoor or two.
As opposed to the current open-source movement, which mostly consists of microserfs beavering away for a salary to make FOSS work better for all of the above?
Let's face it, when the biggest contributor to the Linux kernel is Microsoft, the open source movement has very little to do with RMS' ideals.
This is a weird rejoinder to make in a thread about ESR, who was one of the loudest voices arguing for corporate-friendly language ("open-source" not "free software") and licenses (permissive, not copyleft).
And yet, if that's what it took to drive corporate adoption of open source software, it was almost certainly worth it...
Can you imagine Linux in a world where companies like Google don't contribute kernel patches? Because I think that would look like a world where all the best OSes were proprietary.
One of the big breakout moments for Linux was when IBM added SMP support. Before then it was never useful for anything "serious" or commercial generally speaking.
I used SMP in 1995, using a 1.3 kernel. Board was an Asus with two 100 MHz Pentiums.
I saw about a 35% speedup on a "make -j 3" back then; there was a lot of coarse-grained locking still (BKL), and the hardware wasn't good to begin with.
I'm pretty sure IBM had nothing to do with Linux; the name that comes up in my mind when I think back to that time is "Alan Cox", not "IBM".
I did contract work for infomine.com in 1994; it ran on Linux.
Not necessarily. Remember, the dot com boom happened before Linux was even a serious contender as a server OS; back then Sun Microsystems was king. Google was founded in '98 and the expense of Sun hardware might have slowed Google's profitability slightly but not enough to matter.
We might be talking about the Google-Sun duopoly today instead of the Wintel duopoly today had Linux not existed.
BSD was around, popular, and much more mature than the Linux ecosystem in that early 90s, but was dogged by the BSDi lawsuit. (In fact, SunOS was derived from BSD for most of the 80s, through 1992 or so, when they jumped to AT&T/USL Unix — likely as a result of the lawsuit.) This is arguably the raison d'être of Linux as well[1]:
> If 386BSD had been available when I started on Linux, Linux would probably never had happened.
In 1998, Linux was still behind SunOS/Solaris in maturity and even if it had been even, Intel x86 hardware wasn't remotely on par with what Sun machines could do.
I was using Linux and Solaris in 1998, and doing things on Solaris was a constant pain compared to doing the same things on Linux. (Sun was still trying to push CDE at that point, for crying out loud.) I compiled KDE and Samba for our Sun and dramatically increased its utility.
There were things that Solaris was better at in 1998: big iron things, NUMA things, fault-tolerance things, what we now call "observability". (I still don't have a good equivalent for pstack on Linux!) But Linux was enormously better at many other things, including, most crucially, networking.
That was about the time the Top500 list started morphing into a list of the top 500 Beowulfs.
Ah, but a license, no matter how permissive, does not ipso facto cause the admission of questionable changes into the upstream of a project that everybody trusts.
It just means you can do whatever you want with your copy, within the scope of whatever realm you control.
This might be a good thing then. This has become a pet peeve of mine.
At the last FOSDEM there was a talk on "open source under attack" claiming that anything other than permissive licensing represents an attack on open source:
Two of the three presenters were lawyers for Google and Facebook, two giant surveillance companies with absolutely closed SaaS silo ecosystems. Evidently the "spirit of open source" is free labor for surveillance capitalism and pop-up SaaS products that monetize open source with paywalled APIs and give nothing back.
".. parties who seek to limit the promise of free software" that's an interesting phrase as Free Software has always also been about limiting freedom. After all a license is something that puts certain restrictions on the usage of a piece of software. This is not much different to accusing the GPL of "limiting the promise of free software". Or the AGPL for that matter, which is probably what the OSI proponents are thinking anyway.
This doesn't seem to be accurate, especially as far as OSI and OSD is concerned. Say what you want about the AGPL, but the license does not discriminate against any field of endeavour - it does not state "you can't use this software to offer a network-based endpoint" or anything like that. It just triggers a source distribution requirement in case of such use, by analogy with distributing the software itself to the public (an analogy which is in fact supported by copyright law itself, viz. "public performance" of a copyrighted work). The MongoDB approach post-relicensing is very different.
ESR probably wasn't personally needed. RMS may have been.
Either way, to your statement about "the movement", I think it's important to keep in mind that a lot of what passes for hacker culture today is quite different from where those two hail from. How much overlap is there between the Tao of Programming and modern web development, for example?
Whether that's good, bad, or just a value-neutral observation is pretty much up to the beholder.
ESR talked up his role and theories about how open source worked, but his writing was the main thing that he contributed. His attempts at contributing actual code did not have nearly the same impact. Nor has his description of how open source worked actually had the same impact.
Even the OSI definition owes more to Ian Murdock (the "ian" in "debian") than it does to ESR.
However if RMS and Brian Fox had not single-handedly set about reimplementing the Unix toolsystem back in the mid-1980s, the world would today be different.
As much as RMS annoyed people by saying that Linux should be called GNU/Linux, he had a point. Linus contributed the kernel, but most of what people called Linux at the time was actually written by the FSF.
