Sell hardware, give OS for free. (IBM's early sales model. They didn't do too badly from it, either)
Yep, good, if you're a hardware manufacturer then this is worthwhile for you - essentially, you nuke the whole software business ("cut off its air supply") and gain lock-in/monopoly control of the customer base based on their hardware, not software choices. RMS might not like it, but even RMS hasn't proposed that equipment wants to be free. End results: Programmers who want to eat wind up working for a small number of hardware companies.
Sell Value-Added software and/or hardware. Linux is great, for "general purpose" stuff, but all the mainstream medical and scientific software is written for DOS or Windows. Usually DOS.
Either you keep it closed source, in which case you get nasty letters from RMS, or you make it open source, in which case you get nasty letters from your landlord.
Give the basic software AND the hardware for free (to inflate the Linux market), then sell "popular", non-GPL Linux programs.
Same as before. You've abandoned free software principles by selling closed source software. So why do you want to be on Linux in the first place? If you don't believe in open source/free software, Windows is a better market.
Sell "awkward-to-roll" configurations to specialised markets. eg: SELinux, MOSIX and IPSec will NOT merge, without some serious work. None of the regular distros sell such configurations, so you've no real competition.
Either you find some weasel-argument that lets you get away with not GPLing your mods to GPLed software, and get really nasty letters from RMS, or you're back to arguing with your landlord.
Sell embedded Linux systems, eg: as burglar alarms, automatic pet feeders, automatic oil changers, etc. The more novel the market, the better, especially if Linux' stability can become a selling-point.
But at this point nobody cares that it's Linux. Whatever you're selling, you have to sell on its merits. It's only one step away from, say, the Burlington Coat Factory business - sell apparel and run your cash registers with Linux. You really can't say this is an open source business model.
In the closed source world, a lone programmer can set up shop, develop some software, work at marketing and selling it, and - if everything goes just right - pay the rent and maybe even hire another programmer. The question at hand is: how is this possible in the open source world? Suppose I spend the next six months writing LINABM (which stands for LINABM Is Not A Better Mousetrap). Maybe I rang up some credit card debt, maybe I just burned through a bunch of savings, but either way I really need to see some return on this effort. Suppose it turns out that there are a few hundred thousand people who need LINABM and they are all willing to pay $20 for a copy. Great! But they certainly won't pay a cent if I set my asking price at $0. Closed source means that if I can find a way to communicate to these people that the software exists, maybe I break even, and there's a small chance I might even turn a profit. Open source, in this situation, means I gain hacker cred and some prime real estate in the noosphere, I can walk into any LUG and gain instant friends and admirers, I get all the ego gratification I can stand, maybe I even get some reasonably attractive job offers...but there's that darn landlord after me again.
I know someone who was recently laid off from Nortel. This is exactly the choice he is facing: Get a job, or write software and try to start a business? I've been racking my brains on this, and I can see no possible way for a small firm to launch successfully by releasing GPLed code to the world. As per all of your suggestions above, the only way to make money in an open source environment is to make the money on something else - which doesn't really help you if the business you want to be in is software.
Maybe someone else in this topic will have a brilliant idea, but I've been scouring the Web and I haven't seen any yet...
[ Parent ]