Patents, said Thomas Jefferson, should draw 'a line between the things which are worth to the public the embarrassment of an exclusive patent, and those which are not'. As the value that society places on intellectual property has increased, that line has become murkier--and the cause of some embarrassment, too. Around the world, patent offices are being inundated with applications. In many cases, this represents the extraordinary inventiveness that is occurring in new fields such as the internet, genomics and nanotechnology. But another, less-acceptable reason for the flood is that patent offices have been too lax in granting patents, encouraging many firms to rush to patent as many, often dubious, ideas as possible in an effort to erect legal obstacles to competitors. The result has been a series of messy and expensive court battles, and growing doubts about the effectiveness of patent systems as a spur to innovation, just as their importance should be getting bigger. In 1998 America introduced so-called 'business-method' patents, granting for the first time patent monopolies simply for new ways of doing business, many of which were not so new. This was a mistake. It not only ushered in a wave of new applications, but it is probably inhibiting, rather than encouraging, commercial innovation, which had never received, or needed, legal protection in the past. Europe has not, so far, made the same blunder, but the European Parliament is considering the easing of rules for innovations incorporated in software. This might have a similarly deleterious effect as business-method patents, because many of these have been simply the application of computers to long-established practices. In Japan, fu-ms are winning large numbers of patents with extremely narrow claims, mostly to obfuscate what is new and so to ward off rivals. As more innovation happens in China and India, these problems are likely to spread there as well. There is an urgent need for patent offices to return to first principles. A patent is a government-granted temporary monopoly (patents in most countries are given about 20 years' protection) intended to reward innovators in exchange for a disclosure by the patent holder of how his invention works, thereby encouraging others to further innovation. The qualifying tests for patents are straightforward--that an idea be useful, novel and not obvious. Unfortunately most patent offices, swamped by applications that can run to thousands of pages and confronted by companies wielding teams of lawyers, are no longer applying these tests strictly or reliably. For example, in America, many experts believe that dubious patents abound, such as the notorious one for a 'sealed crustless sandwich'. Of the few patents that are re-examined by the Patent and Trademark Office itself, often after complaints from others, most are invalidated or their claims clipped down. The number of duplicate claims among patents is far too high. What happens in America matters globally, since it is the world's leading patent office, approving about 170,000 patents each year, half of which are granted to foreign applicants. Europe' s patent system is also in a mess in another regard: the quilt of national patent offices and languages means that the cost of obtaining a patent for the entire European Union is too high, a burden in particular on smaller firms and individual inventors. The European Patent Office may award a patent, but the patent holder must then file certified translations at national patent offices to receive protection. Negotiations to simplify this have gone on for over a decade without success. As a start, patent applications should be made public. In most countries they are, but in America this is the case only under certain circumstances, and after 18 months. More openness would encourage rivals to offer the overworked patent office evidence with which to judge whether an application is truly nov