The reasons for the existence of a patent are usually cited as:
If you are curious about the technical details of the patent process, see the general information from the US Patent Office. To quote: The right conferred by the patent grant is, in the language of the statute and of the grant itself, "the right to exclude others from making, using, or selling" the invention. What is granted is not the right to make, use, or sell, but the right to exclude others from making, using, or selling the invention.
Of course, the patent is a double-edged sword. On the one hand, it encourages people and companies to expend the money, time, and effort, to make inventions, hoping to receive some return. On the other hand, a patent gives license to prevent others from using the invention, thereby placing an obstacle to the diffusion of technology.
As a practical matter, patents are often used by corporations for trading. That is, the company does not aspire to exclude others from using its patented inventions, but rather is most interested in making sure it can use any invention relevant to its business. It therefore consciously develops a patent portfolio, so that other companies will want to enter into blanket "cross licensing" agreements with it.
Other companies consciously try to exploit their patent portfolio by actively initiating lawsuits against other companies, hoping to gain significant revenue from penalties imposed by the courts.
Others, including particularly individual inventors, may not have plans to exploit their patents directly, but hope to sell royalty-bearing licenses to companies that will exploit the invention.
Up until fairly recently, patents for "algorithms" were prohibited under a doctrine that basic scientific or mathematical concepts were not patentable. Software was generally considered a embodiment of an algorithm and thus not patentable. This has all changed over the last couple decades or so, and has become quite controversial.
Arguments in favor:
Arguments against:
Finally, it was suggested that possibly a new system of "petite patents" is needed. These patents would be easier to obtain (but more difficult than copyright) and give protection over a shorter period of time.
This takes us off the track somewhat, but viewed in economic terms the patent system is intended to avoid technological inventions becoming so-called "public goods", meaning that they are owned by nobody (or equivalently owned by everybody). The economic problem with public goods is "free riding", where companies (or other economic entities) have little incentive to contribute to (produce) public goods, but every incentive to exploit (consume) them. Without patents, technological inventions become public goods.
This leads us to reflect on the nature of research, using Bell Laboratories as an illustrative example (and a particularly good one).
Prior to World War II, much of "industrial research" was what we might call today "exploratory development" rather than "research". By this we mean that it was directed toward very specific commercial goals, like reducing the cost of telephone service, or providing telephone service across oceans. Following World War II, both industrial and academic engineering research took a turn toward the "curiosity driven" model long practiced by scientists like Heinrich Hertz, who first demonstrated electromagnetic waves, where a foreseeable practical outcome was not necessary. While today we lament the declining support for such research in both academe and industry, in fact this is really a return to an earlier pre-war age when engineering research was much more goal-oriented. Interestingly, at the same time we are witnessing a renaissance of the independent inventor and entrepreneur, particularly in the software industry where the barriers to entry are relatively low.
While the decline in curiosity-driven research may not be lamented by all, there is a danger (and some would say reality) that goal-driven research with longer time-horizons may be slashed along with curiosity driven research. A significant issue here is the increasing tendency to apply traditional financial metrics to the support of research. A financial metric discounts revenues resulting years in the future resulting from research supported today by p^k, where p = 1/(1+r), k= the number of years from the research outcome and the revenue, and r is a prevailing interest rate or "cost of money". Much research of a curiosity or long-time-horizon nature has a long gestation period, and hence a low financial value, although it would be greatly valued in technical or societal terms. This is in spite of the existence of a patent system.
For this reason, much research has to be supported by other means not coupled to such financial criteria. Much of the success of Bell Laboratories up until the 1970's arose from the distinctly different societal-benefit model under which it operated. The research was valued from the perspective of achieving "universal telephone service" though a progressive reduction in the cost of telephony, and was largely paid for not from future revenues, but from an assigned fixed percentage of every telephone bill (called "license revenues"). This provided a stable source of support (another important characteristic of successful long-time-horizon research) divorced from specific coupling to discounted future revenues. Furthermore, after the "consent decree" of 1954, Bell Laboratories was obligated to freely license its technology to the entire industry, making them deliberately public goods. In other words, this successful industrial model of research was actually supported by a broad-based tax, and the benefits were spread widely to the society at large.
Today, we are increasingly decreeing that industrial research that must be justified in conventional financial terms based on the resultant revenues accruing to the organization that finances that research. Since we observe that research results diffuse rather quickly, which results in many research outcomes becoming public goods. (Of course, patents are mechanisms to counter this, with some effectiveness.) Free riding results in inadequate (from a societal view) investment in the research breeding those public goods. As we have moved toward a competitive marketplace for telecommunications, the only viable model for financing curiosity driven or long-time-horizon research is government funding. It is critically important that government continue and expand support of long-time-horizon and curiosity-driven research.