Red Hat, Canonical and MindTouch combined have not had as much impact on computing as Apple or Microsoft alone. I was going to qualify that, but no, I think it's pretty unequivocal.
Also, your selection of companies seems a bit deceptive. These companies' profit strategies are very narrowly focused on the one market where it's possible to make money with free software — and even that isn't really making money from the software. You have to be in a market where people will pay through the nose for support contracts, because the software itself isn't making much money.
The one real exception I can think of is Mozilla, which makes decent money off of Firefox. Otherwise, unfortunately, it does generally seem to be necessary to bind the users enough to make them pay you for the software. (That's not a sarcastic "unfortunately" — I do think it would be great if I could reasonably sell free software as profitably as non-free software, but I don't see any evidence that it's feasible.)
> Red Hat, Canonical and MindTouch combined have not had as much impact on computing as Apple or Microsoft alone.
Wouldn't that be a question of time and business model? Apple is a hardware company and both were around when the PC (not the IBM one) was invented.
> isn't really making money from the software
It's very hard to directly make money from software sales if you grant your customers the right to share it. Software companies make money out of the scarcity of software ("You want Windows? Ask Microsoft"). Free software is naturally abundant. It's possible, but not easy, to make your users pay directly for it.
And Mozilla can't ask users to pay for it. Not only because it's infinitely abundant (anyone can give you one) but Microsoft's IE comes bundled in the OS most Firefox users run. Opera does that and look at their market share.
And that is why it is generally necessary to "imprison" (as Stallman would say) the user. Otherwise you're making a product that won't make money. There are a few markets where you might find a connected business to recoup the money you spent on software, but it's not a general business model for development.
That's what I was getting at: It's hard to be rewarded for free software, so there really is a dichotomy between free and commercial software. It's not just Jobs and friends being dicks and restricting customers.
No. You are only making software that doesn't usually get sold. Most of it isn't, but you can custom-develop - building upon piles of other free software - and then license it under a free license to your client who chooses whether to share it (always under a free license) or not. Free software can be sold. I know this because I did it a lot.
> It's hard to be rewarded for free software
OTOH, it's much cheaper and easier to develop it. With closed software, you have to invent your own wheels. With free and open-source software, you are free to use the ramjet engines other people have developed.
Again, you're outlining a highly specialized business plan for custom development — not a general principle for making money with FOSS. If I freely license my next game, I'll lose my shirt. If I create an awesome, game-changing productivity suite and GPL it, I'll lose my shirt. If I do anything outside of some very tiny niches, I'll lose my shirt.
There are a few small areas where FOSS is as commercially viable as anything else, and you'll notice those are also generally the areas where it seems to be most mature. You can't generalize those tiny areas and say that making open-source software in general is a reliable income source.
It's not possible to make a business out of selling goods that are infinitely abundant. Free software is one such thing.
Making open-source is not a direct income driver and, apart from very limited niches, will never be, but, nevertheless, it allows a business to have full control of its technology stack.
Company A builds a great game using a closed-source library built by company B and company C builds another great game using, say, an LGPL library. When the underlying platform changes, breaking both libraries, company C is not subject to whatever the strategy of company B is and, therefore, can be first to market with a new, compatible, version.
This is a hypothetical situation, of course, but illustrates one important thing about free software - most people don't really make it: they use it and, from time to time, and, if and when the need arises, add a little improvement here or a fix there. Its strength lies in the sheer number of people that give a little hand, scratching one another's itch.
>There are a few markets where you might find a connected business to recoup the money you spent on software, but it's not a general business model for development.
Also, your selection of companies seems a bit deceptive. These companies' profit strategies are very narrowly focused on the one market where it's possible to make money with free software — and even that isn't really making money from the software. You have to be in a market where people will pay through the nose for support contracts, because the software itself isn't making much money.
The one real exception I can think of is Mozilla, which makes decent money off of Firefox. Otherwise, unfortunately, it does generally seem to be necessary to bind the users enough to make them pay you for the software. (That's not a sarcastic "unfortunately" — I do think it would be great if I could reasonably sell free software as profitably as non-free software, but I don't see any evidence that it's feasible.)