> The patent system just lets people hold society to ransom: it's the opposite extreme to the person who refuses to disclose how they did something and who takes that knowledge to the grave. We need to find (or rediscover) other incentives that help society to progress as rapidly as possible without needless duplication of effort and without needlessly obstructing access to humanity's shared knowledge.
The question is how high is the probability that someone else would invent the same thing.
If it's a very high probability, then patents are bad (the "hold society to ransom" part)
If it's a very low probability, then it's the "takes that knowledge to the grave" issue and patents are a very good deal for society. The knowledge is available, just after a 'short' timeframe.
In the drug industry, the investment to get a drug approved is very high, chances success for any individual drug are low, so it's arguable that companies will not go through all this expense if they don't have some way to be confident that when they _do_ find something that works, they can earn the money back.
In the software field, the cost of inventing something is _very_ low, and it's almost always done as part of a project that has other benefits to the inventor. As a result, the invention would be done with, or without, a patent at the end of it.