I think the HN community's approach is wrong: patent abolishment isn't optimal, heavy patent reform is. Patent terms should be regulated and adapted in each industry like the Fed Funds Rate, not like the 10 commandments. Software is definitely an example of where they should be abolished, though.
This is probably something the FDA should do anyway. Generic drug makers can look at a drug's signature, but there are so many other variables that can affect how a drug performs (quality control of the factory, quality of ingredients, recipe, procedure, etc.) that the FDA should require generic makers to go through the same testing process because they're different drugs.
The generic drug makers will have less overall expenses to bring a new drug to market (because someone already figured out a particular protein sequence that works in XYZ manner), but the original inventor will have time to milk the market as the first-to-market, encouraging innovation, and we can still get rid of our awful mess of a patent system.
Generally if you've built up a whole system (like the patent system), and it's really only beneficial for a select few (pharma), then maybe it'd be better to design a different system just for them (like the one proposed above) and don't make everyone else suffer the consequences.
It's too easy to say it this way. Any data to show us why the case of the pharma is so different from others ? The costs involved in the pharm industry depend heavily on: - government regulation - expected returns
It's not evolving in any kind of "free market" at all. It's a very poor example to use to justify patents.
Open Source Drug Discovery - http://www.osdd.net/
Research article on PLoS; "A Kernel for Open Source Drug Discovery in Tropical Diseases" - http://www.plosntds.org/article/info%3Adoi%2F10.1371%2Fjourn...
Also of relevance, the UK patent on using a chemical that gives you an erection as a cure for impotence was thrown out for being obvious. But the patents on how to mass manufacture it are still in effect.