Before I start I'd like to comment on NetPresenz 4.1, which I use for running my web site. I just found out that the price is now $75 per license. Wow. I don't know when that changed, but I bought my license for $25 or $35 and before that it was $10, and this while it remains at version 4.1. Not that I begrudge Stairways trying to make an income, but perhaps there is another way?

Which brings me to the main point of this article. I just read Eric S. Raymond's third paper on Open Source software: The Magic Cauldron, which I highly recommend along with the other two parts of the trilogy. Not quite as polished as the previous two, it still makes a somewhat compelling case for how to make money in an Open Source world.

I guess it shouldn't be surprising that the big income in the software industry is not in making software, but in supporting it. Although that's not the model on the consumer side, that matters less because the business side is still where most of the revenue comes from. Case in point, when West Coast was looking at what Oracle software to buy, license costs made up about 25% of the expected cost to buy, install, train, and support the software.

What do companies achieve by Open Sourcing software? Quicker and better bug fixing since there will be more developers looking at and fixing the code. More features and faster development. The goodwill of a bunch of smart people, some of which make buying decisions (for your other products). The first is the most important if you provide free or low cost support, because you can't afford to have your programmers fixing code when they could be writing new code.

The big disadvantage is not being able to exclusively sell the code and that competitors have access to your code. The second is not as important because GPL licenses prevent your competitors from using your code and selling it as part of their exclusive product -- anything they do you can also copy. The first point is more tricky. At the beginning of a product life cycle it is advantageous to sell it exclusively, as you have a monopoly. Near the end it is better to Open Source it to reduce your support costs and because someone else can preempt you by Open Sourcing their code and getting developer loyalty.

There are a couple of other models of Open Source that work, mostly for non-software only companies. One is the shared-development model. Companies that want a good solution yet don't have the budget can contribute their programmers to a project that creates a solution for everyone. Especially appropriate if your revenue stream doesn't depend on the product being made. This is how the Apache Project was started.

The other model is the shared-support model. You have your programmers write an in-house solution, but you don't want to maintain it, especially once those programmers leave. If you Open Source the project, other people use it, they support it, and you don't worry about having legacy code that is not maintained. Cisco did this for some network printing solution they used internally.

What is also starting to happen is that companies that benefit from Open Source are hiring developers to help maintain and expand code. O'Reilly does this, as they make their money on documentation; RedHat also, because they're the Linux market leader so improvements in Linux will benefit them the most. This sort of paying for works that will be publicly freely available is reminiscent of the Renaissance, when cities and rich men hired artists to produce works. It increases good will for that company, and people do notice that.

So after reading Raymond's three pieces, I think that Open Source has a big future ahead of it. Its current incarnation in Linux, which more "open" than the older BSD and FSF software, surprisingly seems to work, as Linux software growth is astonishing. I don't think databases and ERP applications are at the point where they need to be Open Sourced, so I think my company is fine for now. But the time will come when the software we write doesn't make any money initially, but sets up the rest of the revenue chain.