As you will note, I've previously argued with you against software patents. Since our last discussion I've talked to a number of developers and no one can recall any educational classes or books on software development that even suggested checking to see if your code infringed on any patents as you were writing it. None of us have a clue on how you would do that with any signficant size software.

I think it's more like writing a novel than building a device. Would you patent "boy meets girl" as a theme, or would it be patented if the boy was from a poor family and the girl wealthy? Stories, like software are built of a collection of common themes that are used to geneate a whole.

Hi dbhalling, thanks for your detailed and thoughtful reply. It is the prohibitions on tinkering, modifying, extending and improving which originally drove me into a strong open source position.

Another driver for me in this was the vibrancy of the open source community. The open source movement has dramatically lowered the entry cost for new players, which has stimulated huge creative effort. Bill Gates became the richest man in the world largely from his astute decision to add open source networking software (the BSD Unix TCP/IP stack) to Microsoft's Windows operating system. Today, more than half the world's websites rely on open source software (Apache server). More and more cellphones rely on open source software (Android phones, using the Linux operating system kernel).

And there's another aspect of open source which would delight the Gulchers - the ability to audit the code and discover if some looters have planted surveillance backdoors in there!

Anyway, here's one of my biggest frustrations as a developer: software that does not allow itself to be interfaced with other software. I find your suggestion interesting - that stronger patent protections might have had a more liberating effect on software, by removing the need for software publishers to cripple modification and interoperability.

This still leaves the problem of patent discovery, however. There is a mass of software patents written up into vague, confusing, needlessly abstract terminology, which makes it difficult, time-consuming and expensive for a developer to find out if the code they're writing actually infringes on someone's patent.

I wonder - could there be a regime which maintains the inter-operability, transparency and low entry barriers of open source, while maintaining intellectual property rights which are cheap and easy to stake out, defend and (in case of others' IP) discover and uphold?

There is a concept in patent law called the first sale doctrine. It states that if you buy a patented item then you have a right to use, repair, modify, combine the purchased item in any way you want. The only limitation is “completely” rebuilding the invention. I believe there is a somewhat similar concept in copyrights, but it is more limited.

This problem stems from regressives screwing up contract law (bear with me it’s a long story). Some law professors though contract law needed to be updated and so they created the UCC which applies to certain types of contractual transactions. One of the requirements is an implied warranty of merchantability (meaning the warrant was required regardless of what the contract said). Unfortunately, in the early days of standalone s/w the products could not meet this warranty. As a result, they decide to license the s/w rather than sell it. This took them outside of the UCC and the implied warrant. Because the s/w is licensed not sold then it avoids the first sale doctrine. Also because the ability to enforce and obtain patents directed to software inventions was limited, s/w companies relied on their copyrights (which were weak), and their licensing agreements. As a result, their licensing agreements stated you would not reverse engineer, modify, or even look at the source code (of course they never provided it). However, I do not believe the auto companies sell you the car and license you the software. So I don’t think they have a leg to stand on under standard copyright law. But the DMCA may apply if the s/w is encrypted.

The UCC and weak or non-existent patent protection for s/w in the early years caused a number of other problems. Including that when you got the s/w you could not modify or integrate it into other s/w. This would be like buying a car and not being able to put different tires or a different muffler on it. This clearly frustrated many who were sophisticated in s/w which started another bad solution – radical open source (non-radical open source makes sense for many things – think of standardized interfaces). This would not have happened if there had been strong patent protection for s/w implemented inventions and the UCC did not have the implied warrant of merchantability.
So once again we see one idiotic do good thing by regressives over 50 years ago and one anti-property move by regressives over 40 years ago compounded into numerous problems – AND the solution is more patches rather than solving the underlying problems.