Linux and the open source movement pretty much disproves that theory. Better (or "at least as good") quality kernels than the two biggest software companies and biggest defenders of intellectual property rights on the planet could manage.

We should really look at the evidence rather than repeating the mantra.

I don't know if Linux/Open Source/Free Software disproves the theory, or actually proves it.

The argument would be: Microsoft and Apple feel that existing IP protections on software would not sufficiently prevent competitors or consumers from 'stealing' the results of their work. So, they don't release the source code. Because they don't release the source code, there's a duplication of effort between Microsoft, Apple, and any other kernel-writers.

If there existed extremely strong IP protection laws for software, that required release of source (like a patent, basically), then Apple and Microsoft might release source, duplicate less effort, and their disadvantage to Linux would be reduced.

Open Source is what happens when you have a system that encourages these secrets to be shared. The only difference is, Linux manages to get R&D effort without the benefit of a monopoly-grant, or indeed, without any direct profit motivation.

Apple didn't manage to build a decent kernel themselves so took a "free" one. Nothing to do with stealing their work. You are aware that Apple uses an open source kernel?

First, I was trying to relate the ideas behind the patent system not to describe how it works in practice. So, while I always like to talk about how gnu and Linux took over the world, it's not relavent to the discussion of rents vis a vie patents.

Second, Unix is a bad example because AT&T never tried to patent it. I'll ask my dad about it. He worked at bell labs during this time (he has about 2 dozen patents and an IEEE medal). I think the attitude around there was that software wasn't something you could patent. My dads research was in lasers so I heard much more about lightwave devices than software growing up.

I think most times its worth the "redundancy" in cost of multiple parties discovering the same thing, because then at least they can use it. The way things stand now, plenty of biotech companies just shelve patented things and then its no good to anyone. This can't be easily remedied by "adjusting" the laws, since the lawmakers will always and quite practically be behind the technology, and sometimes not ever understand the technology or consequences.

Either way there is a cost to use: either paying for the rights to the patent, "paying" in completely arbitrary time determined by patent laws which do very little to distinguish completely different technologies and lifespans, or paying to reinvent in trade secret scenarios.

I feel like with trade secrets its much more in our control to remedy situations where many people need access to a technology quickly. It allows for dynamic solutions. For example, private contracts could handle the patent use case without necessarily having IP. They could provide the end product as a service, or provide know-how to build the product under a "subscription-service" where the client agrees to pay a very large fee over the course of several years/dependent on unit sales/etc (the analog to the royalties/licensing fees), with the punishment being spelled out in the contract that if the tech is replicated in some other way or leaked, the client company must pay a large "breakup" fee. That way the client is incentivized to use the tech and keep it secret as well. Client companies can then assess whether these fees are better than the cost of R&D they would have to undertake to develop it themselves. Just look at software: there's plenty of stuff I could build myself, but I happily pay for an existing solution despite not being "forced" to. I think we very early on decided on a patent model and just threw up our hands and said 'no one would possibly invent otherwise!'

Its also worth noting that most research is publicly funded (scientists in labs discovering things) while most development is done by large companies (putting drugs through clinical trials, and bringing them to market). Certainly in the biomedical field to which my work is related.

That can't possibly work though. Patents only give an incentive to patent things that can be expected to be independently discovered or reverse engineered sooner than the patent expires. There is no reason to patent things that could be kept as trade secrets for longer.

Sorry, I didn't mean to imply that it worked, only that this was the idea behind the system. The practice of our patent system is pretty far off from this ideal.