Take the pharmaceutical companies as an example of where patents are perhaps worthwhile:
- It costs a lot of money to produce a new drug; high cost == high risk == unattractive field for investment. Patents lower the risk because the pharmaceutical companies know that they will be able to recoup their investment if they actually produce something useful.
- Likewise, it costs a lot of time to do the research and produce something that works. The 20-year timescale (http://law.jrank.org/pages/9086/Patents-Patent-Duration.html) of the patent is approximately in line with the effort expended.
- It is comparatively easy to tell if a drug works.
- After the patent expires, the drug is probably still useful...e.g. aspirin hasn't stopped being a good painkiller just because it's a commodity.
Compare this to software patents. None of the points match up:
- Production costs are zero == low risk == attractive field for investment.
- Production time is almost zero. You can submit software patents literally as fast as you can write them up.
- Most software patents are completely intangible and have very fuzzy edges. If you're a patent examiner in 1980, how do you verify that this "web browser" idea will in fact work?
- After 20 years, the patent is likely to be worthless. As a trivial example: yes, people still use GIF but JPEG and PNG are far more common.
All of these things encourage the system we have now--"throw it against the wall", patent trolling, etc--instead of what patents were Constitutionally intended to be: an incentive to grow the arts, sciences, and economy.