sure. But I think 99.9% of people just don't care if they're "free" by his definition of free.
If I buy a laptop, I consider myself free, even though I don't have the schematics, the moulds used to make the plastic, the development details etc.
But what if the laptop co goes out of business, and my laptop case gets a chip! How will I make a replacement case if I don't have access to the moulding details etc! Well, I don't really care that much.
It's just an odd thing to strive for IMHO.
I think one of the problems with open source is that it assumes anyone can just 'take over' some source code and fix bugs etc. The reality however is that it's not that simple.
Just like giving me plastic moulds is nice in theory, but pretty useless in practice.
When I use copyleft software, this rarely happens. I can only assume it's because some other user has experienced the same annoyance and fixed it already. But when I do run into something that's broken, I am able to look at the source, make a patch, and submitted it back upstream. This is usually with software that has a smaller user base.
I value this freedom and am glad that RMS and others have worked so hard to spread it.
The only important thing here is having people behind a project that care, respond to any queries, and are actively developing/fixing/maintaining.
Good software isn't good because it's open or closed source. It's good because people are sitting behind it and care about it.
That's personally all I care about - that the people 'behind' the software care, and will fix any issues that need fixing.
EDIT: Clarification (apparently, the above is cryptic enough to warrant a downmod): in any software project, the founders of the project can quit, go broke, die, or just be much less active. And no one can guarantee that it won't happen.
If the software is GPL, you have the guarantee that when that happens, some more developers can jump in and take the lead. If no one do so that means only one thing: no one cares enough (including users, which could hire someone). Good riddance. Example: CVS is dying because everyone is willing to switch to SVN.
If the software is proprietary, that possibility may be denied by the original founders. Which means a proprietary software can die while people still care. Example: old versions of Windows or Word, or games on old platforms.
Which leads to the obvious question: for any proprietary software, can you guarantee that it won't die before no one cares? I think not.
Usually some aspect of monetization.
For instance, plentyoffish.com is free, closed source, but has someone behind it who cares, because it's his livelihood.
Your examples make no sense: Windows still exists and most games are played one or twice. Sure, old versions are not available, anymore, but who want to run them, anyway? The same holds true for old versions of free software: They may be still available somewhere but who wants to run them, anyway?
While you're talking about dieing software, what about creating software? Can you "guarantee" that a particular free software will be created if a large number of people want it? I think not.
Proprietary software, on the other hand, comes close to a guarantee if there's a chance of it being in demand.
This is why sociaty needs both -- and why Mister Stallman is wrong to say, only "free" software should exist.
My examples makes no sense to you, but they make sense to me: I was thinking about a particular Windows95 game which I would have loved to see ported on the Pandora [1] : Earth siege 2. This is not possible because there is no source code to port, and emulation is out of question (the Pandora isn't powerful enough). The only reasonable solution I have is to clone it, almost from scratch.
About creating software, I agree that some (many?) software couldn't have existed as free software (only). However, this is not a proof that society needs them[2]. Even then, this does not prove Stallman wrong: he cares about freedom. Availability and quality are secondary to him.
[1]: www.openpandora.org
[2]: By the way, if society does need them, then I think the government should write it or finance it.
The problem with software is that once it is integrated into your software ecosystem you often wind up locked in. If that software stops being able to work for you (for instance your license stops working and you can't purchase another one), then you frequently have a very difficult transition ahead since there is no drop in replacement available.
RMS sees this, knows that users often don't realize their potential future pain until they are committed to it, and concludes that non-free software is evil. By contrast an economist sees this, assumes that the market will be rational, and writes a book like Information Rules explaining how one should best act given the economic facts.
My view is that most people who really care about software have been around the block enough times to have experienced the economic fact and in practice act more like economic theory says you should act. Of course this provides an economic argument for free software, which has of course lead to the open source set of arguments.
YMMV. And on this topic it tends to vary widely.
So it kind of hearkens back to the day when computer users were also competent programmers.
I think the modern day analogy to proprietary software would be someone sending you a locked excel sheet with a license that says 'you may not change, copy or distribute this excel sheet,' and then having glaring error in the spreadsheet.
You see the error and want to fix it but are obligated not to because of some silly rules. Instead you have to email the idiot who wrote the broken excel sheet, and have him fix it. (And wait for him to reply, which he may or may not do).
This resonates with me because I hate wasting my time on extra steps. I should be able to email the guy 'Hey, i fixed a bug, look at this' and be done with it.
Granted, the majority of computer users now are not of this personality type, and even if they are, wouldn't have the expertise or experience.
For FSF to really win, we would at least need a much higher level of computer education in the general population. (IMO 90% of people are functionally illiterate, with regard to computing).