You cannot really blame people for preferring something free to something expensive.
I can, if there is a significant difference in quality. Here, we are talking about tools, whose quality will immediately affect the quality of the result.
I obviously understand the reason why hobbyists prefer free over nonfree; I'm a hobbyist myself. But, when doing a paid work task, I do pay for proper tools.
But this would destroy the mechanism which forces the car manufacturers to produce better and cheaper cars.
The
current mechanism which forces the car manufacturers to produce better and cheaper cars.
Fact is, all industries have gone through a fundamental revolution changing business models every couple of generations for a couple of centuries now. The answer is not to freeze the business models, but to develop them so that they match the market, and the overall market drives what the cultures involved want and need.
Let's say that what you outlined actually happened, and suddenly cars were a commodity instead of a luxury item. How does that differ from how internet came to be?
Would you have stopped the development of the internet just to protect the business model of telecom and cable TV companies?
The fast and hard fact is that companies are not human, and are not worth protecting. It is the
market, the competition, that keeps businesses healthy, and that is what we (whoever has the executive power) must enforce and direct so that overall, the market does what we want.
A major problem in current software choices is that ever since Microsoft gained dominance, humans have accepted that software is unreliable; that that is the nature of software. There are a few niches (like medical devices) that are supposed to have different values, but experience (their failure rates and typical bugs) shows that is simply not true, they're just as crap as every other piece of software. That attitude, that software is fundamentally unreliable, is pervasive throughout human computer use: we're basically programmed for that now. Yet, Microsofts products are nonfree, and cost money, even though better free alternatives exist.
More importantly, the shift even in free software is from reliability and robustness to ... I dunno, surface bling? marketing? catering to lowest common denominator? Whatever the heck systemd is; all I know that it is neither KISS nor UNIXy. Fewer and fewer people are interested in OpenBSD or FreeBSD. Crappy and buggy Linux distributions are most popular; they don't need to be robust or reliable anymore. Hell, they've brought back the "you need to reboot your computer now" idiom!
If the mechanism is destroyed (as in Soviet Union, or in software market) people stop producing and everything goes down the drain.
No, that happened in Soviet Union because competition was eradicated.
In the software market, competition is weak because people accept crap from
both commercial and free software. There is no demand for anything better. You claim it is because no-cost software saturated the market; I claim it is because people are easily satisfied with bread and circuses, even when something better is available.
Here in Finland, it is completely accepted that multi-million IT projects often simply "fail", producing no results. They are blamed on some of the lower-level workers, the bosses get their bonuses and switch to new projects without any harm to their reputation. This is accepted practice, and is also the reason why IT product quality in Finland has gone to crap.
Let me go back a couple of decades, when I first started running a Linux-based mail and file servers for a small department in a Finnish university.
This was the era of Sendmail, a buggy insecure piece of shit that everyone was using, because it was what everyone was using. Me, I've always had a paranoid streak, so after testing a few different options (with Exim, procmail, and qmail in the top three I chose from), settled to qmail, because of its security approach.
At every single turn, I had to explain my choice to people who could not themselves explain why they used Sendmail and not something else.
The biggest issue I ever had with it, was that it was too fast in processing internal mailing lists, confusing the main University Sendmail installation; limiting the rate at the Sendmail end fixed that. After a couple of years, even the main Uni server admins came to accept it as a reliable, although "quirky" choice.
The exact same situation was, and remains, with Bind. Having fed up with bind security issues, I switched to djbdns (for a local NAT).
Fast forward to today.
Exim has the same architecture as Sendmail, and although is much better security-wise, is Swiss cheese compared to qmail. Most people still use bind, and simply live with the frequent security issues. qmail is a tiny niche. Reliability and efficiency wise we have gone backwards, not forwards. "Patch Tuesday" is a fact of life, and not a joke anymore.
What happened, was that combined with the increasing memory, bandwidth, and processing power, we reached the satisfy-the-humans level with crappy software. It does not matter whether it is paid or no-cost, free or proprietary.
So yeah, I definitely blame us humans for having such low expectations and being so shortsighted. Stupid, really.
Many otherwise good tools in the embedded space - compilers, RTOSes, GUI libraries, etc. - have become a huge liability for people by suddenly disappearing from the general market when one player (e.g. a silicon vendor) bought them.
I know nothing of the RTOS market, but single-vendor products freak the shit out of me. I've been burned too many times. I don't trust them at all.