How much do open source license terms matter?

Where you stand on the open source incline, whether it's near the top with a Microsoft license, at the bottom with the GPL, or somewhere in the middle with a BSD license, is supposed to matter.

But, almost unspoken in Matt Asay's post today is a key point. These claims are, on the whole, untested. (Armin Shimerman, shown here as Quark, also appeared in 2005's "What the Bleep Do We Know?")

That's why wise heads in the blogosphere are spending time debating a case involving, literally, an Artistic License on model trains.

In this case it's the nature of the license, and its basic enforceability, which appears to be at issue. There is, as yet, very little case law on the mixing of licenses, or on different terms used in different open source licenses.

The reason is that all open source follows what Trekkies will recognize as the first Ferengi Law of Acquisition. "Once you have their money, never give it back." For money, read code.

Open source overturns the trust relationship which existed in software for a generation. The old rule was you took their money while giving people nothing and guaranteeing them nothing.

It's not a sale, it's a license, a permission to use the tool.

In open source, customers get the code, even before they give the writer a dime. They get the trust, not the vendor. Michael Tiemann of the OSI might call this "libertarianism in action."

In practice the relief available to an open source vendor is limited by the fact that the customer has the code. You're trusting them to treat you fairly, under the terms of the license, but the chances any dispute will wind up in court is small.

It's a tough business, basing your future on trust in your customer. Surprising how many people have the lobes for it.