> "The purpose of patents is to get things invented that would not otherwise be invented"
The purpose of patents is to advance public knowledge. It was understood that people would invent regardless (as people always had). But without patents inventors obfuscated, sealed devices, etc. And when they died, chunks of knowledge often died with them.
But inventors were reluctant to make their secrets publicly available for everyone to see and possibly copy. Hence the temporary monopoly was extended, not to make their job in the market easier, but as payment for sharing their knowledge with the public.
> The purpose of patents is to advance public knowledge.
IIRC, the primary purpose of patents, since they were invented in 15th century Venice, has been to eliminate the monopoly of trade guilds. This was an issue well into the 20th century. But has software ever had this problem?
> "has been to eliminate the monopoly of trade guilds"
Exactly. By making their knowledge public.
> "But has software ever had this problem?"
While the history of software looks nothing like Renaissance Italy, the future may -- particularly in a hypothetical world without software patents.
Historically, the root of even arguably-patentable software is expressly unpatentable math that has largely been developed by publicly-funded institutions. So that even when, say, the Fraunhofer Institute was granted patents over parts of the mp3 codec, even absent the patent disclosure, it drew attention to the root math and gave rise to alternatives.
But looking forward the situation is more worrisome. As public funding has failed to keep pace, only the potential to patent research has brought in the private funds that have continued to drive innovation. Without that patent potential, there'd be little reason for anything beyond the hint of promising research to be shared publicly. It would be pursued privately, without published articles, without any public disclosure on the other end.
And given the trend for modern software research to manifest as internet-accessible services, it doesn't seem like there would even be an opportunity to reverse-engineer or decompile the ultimate products that such research generated. To say nothing of the rapidly increasing complexity of modern software and the increasing rate at which it's tied to custom silicon.
Consider an example: How in the world would someone be able to reverse engineer Siri in a world without software patents or the potential to patent publicly-funded core research?
The public has no access to the server-side source. We have very limited access to the client-side source. We have even less access to the client-side custom silicon [1]. Even if it were plausible for a not-for-profit effort to untangle a project of that magnitude, they simply do not have the access necessary.
[1] Though client side silicon is less integral in the current implementation, it's no stretch to imagine that Apple will continue to refine and accelerate its algorithms through heavier use of custom DSPs. Nor is it news to say that it approaches the implausible for not-for-profit efforts to reverse engineer modern silicon. And the relentless pace of fabrication promises to only make things more difficult going forward.
The purpose of patents is to advance public knowledge. It was understood that people would invent regardless (as people always had). But without patents inventors obfuscated, sealed devices, etc. And when they died, chunks of knowledge often died with them.
But inventors were reluctant to make their secrets publicly available for everyone to see and possibly copy. Hence the temporary monopoly was extended, not to make their job in the market easier, but as payment for sharing their knowledge with the public.