Unfortunately, I can't find a link with this claim to share.
If RMS made a choice to use proprietary software, with the only alternative being his dead, there is absolutely no moral conflict, or even anything against his principle. The only issue could only ever arises if he claimed "Free Software is more important than human lives", which I am pretty sure not a claim he, or anyone, has ever made.
It's basically the same principle as if you point a gun to my head and tell me to kill a puppy or get shot (and I'm certain you would shoot me). It's obvious which choice I would make, and I hardly see that as an evidence of my hatred for puppy.
Valuing freedom (even in limited spheres) more than human lives is a pretty familiar American idea. For example, recall Patrick Henry's "give me liberty or give me death" and consider that he was talking about the prospect of being taxed without his consent, not enslavement. Also, recall Private Eightball from Full Metal Jacket:
"Personally, I think, uh... they don't really want to be involved in this war. You know, I mean... they sort of took away our freedom and gave it to the, to the gookers, you know. But they don't want it. They'd rather be alive than free, I guess. Poor dumb bastards."
That does not answer your question though. It goes both ways: if the doctor recognises that a firmware tweak may save your life, then she should be able to do so.
The answer is obvious based on your own personal principles, or based on RMS's stated principles?
>The only issue could only ever arises if he claimed "Free Software is more important than human lives"
He claimed that it is categorically not good to use non-Free software, with one exception for using non-Free software to develop a replacement for that software; and that "we must resist stretching [that exception] any further": https://www.gnu.org/philosophy/is-ever-good-use-nonfree-prog...
That said - RMS has iirc said in the past that for medical emergencies he's willing to make a personal exception on this stance.
[0]: Medical devices can't be GPL compliant due to the anti-Tivo clause combined with regulations of the FDA that demand that a medical device will always behave the exact same way in the exact same situation. This is also extended to the software, meaning that not being allowed to reflash the software if you're not the manufacturer is a requirement.
What is the anti-tivo clause?
--
Anyway, TiVo wanted to get onto the GPLv2 train like a lot of companies, but they also wanted people owning their devices to be unable to partially modify their firmware even as they distributed their copies of the firmware onto TiVo boxes. They enforced this by having a digital signature check that only allowed TiVos own software to run if the software stack was entirely TiVo. It would also work if the stack was entirely FOSS because only TiVo checked it.
For the sake of clarity; Bradley Kuhn of the SFC iirc investigated the TiVo boxes at the time, the signature checks only were an issue if you partially replaced TiVos software, not if you went the whole hog on replacing it; excercising GPLv2 rights was completely possible and you could turn a TiVo box into a XBMC (these days that's Kodi) box. You just couldn't run TiVos software ever again if you did that thanks to a hardware check.
The FSF took personal offense to this practice and dubbed it "Tivoization", which is a shorthand for "attempting to restrict the ability to install or link with Free Software via hardware DRM". (Which yes, is a modification of what TiVo actually did but really that part is just par for the course with the FSF.)
As a result, the GPLv3 includes an explicit clause that if your software is distributed in a non-code form, that you also must distribute all the information required to be able to modify that distributed software, if there's parts on the device that prevent you from doing so. (Or in plain terms: on restricted hardware, you must give up the signing keys if you preload the hardware with FOSS software.)
--
This clause is usually just called the anti-Tivo(ization) clause and it's... pretty damn controversial. It's the main reason why the Kernel is still GPLv2 and not GPLv3; Linus Torvalds personally considers this clause to be a significant enough alteration of the "deal" that FOSS provides for the kernel, so he didn't upgrade (which to be clear would've also been very difficult since the kernel is 2.0-only, not or-later, so he'd require approval from all significant contributors at that point). It's also often cited as the main reason why the GPLv2 has/had strong corporate backing but software licensed under the GPLv3 has always kinda had issues with that; the v2 was seen as more "fair" in that companies were willing to work with its terms, while the v3 was seen as basically forcing them to give up important parts of their trade secrets and made them antsy of working with the FSF in general.
And as mentioned before, in some fields (medical is the one I know of, but I'm pretty sure there's a few others), regulatory compliance is impossible with v3 while very much possible with v2 because of this clause.
That article was written before the finished GPLv3 was published, they had only published a stricter draft.
Therefore; I'm not wrong.
+ the opposition against proprietary software does not mean that using proprietary software is immoral on its own. (Although you'll find plenty of people who will make this some sort of purity contest, rms is not one of them.) The moral evil that the free software movement address is to be forced or coerced to use proprietary software. The mere use of proprietary software only becomes a moral issue when it results in drowning out free software. Much like the widespread acceptance of proprietary mobile phone apps leads to cultural shifts that essentially force everyone to use said apps to participate in life. (See wechat in China.)
+ The second is addressed here: https://www.gnu.org/philosophy//po/who-does-that-server-real...
As others have pointed out, there is no moral dilemma here. The demand for computational autonomy is not a dogmatic, religious belief that conflicts with medical treatment of a disease.
FSF gave Karen Sandler an reward for her work on this: https://sfconservancy.org/news/2018/mar/26/sandler-fsf-free-...
"Richard Stallman, President of the FSF, presented Sandler with the award during a ceremony. Stallman highlighted Sandler's dedication to software freedom. Stallman told the crowd that Sandler's “vivid warning about backdoored nonfree software in implanted medical devices has brought the issue home to people who never wrote a line of code. Her efforts, usually not in the public eye, to provide pro bono legal advice to free software organizations and [with Software Freedom Conservancy] to organize infrastructure for free software projects and copyleft defense, have been equally helpful.”."
Linux is not a GNU project. But it's free software, so I suppose RMS could use it without major pain. Although he'd probably prefer Linux to use GPLv3.
He's not unreasonable, he's just very adamant about software freedom.
I know it's not foss, and I'm aware of mach, but QNX still holds a place in my heart.
But things are changing and linux is getting some traction.
The healthcare industry is very slow moving (as it should be).
Results in a lot of stuff getting rejected for FSF certification despite being just as open practically as the certified stuff.
A car probably just isn't seen as software to him if it's just the internal control systems. Maybe the entertainment system these days would be.
To my knowledge the FSF doesn't consider firmware blobs with no way to flash "part of certification" because it would be inoractical to do so. By definition "software you can't change" isn't "free"(as in speech, not beer), so I don't think it's an ideological line.
So they either have to consider any device with unmodifiable software to be non-free(basically removing entire device classes), or be realistic and draw the line somewhere. "Can you modify the software on the device from the rest of the gnu software" is a line that matches up with their ideals fairly well IMO.
Basically, Richard Stallman is well aware that cars are bundles of non-free software(see: https://jalopnik.com/richard-stallman-weighs-in-on-the-check...). He has voiced opinions that software controlling the brakes in your car should be FSF and not binary blobs that come pre-installed with no visibility.
You can search his site for the word "ticket" to get an idea of how he feels about software systems in public transport.