Well, first, as almost everyone involved in the patent reform debate acknowledges, we must fully fund the PTO.
...There are not nearly enough examiners to review, revise, and recognize incoming patent applications. There is a backlog of approximately 400,000 such applications -- inventions just sitting around, gathering dust, preventing their inventors from exploiting the fruits of their labor. For the typical application, two years elapse from filing to issuance, a period expected to rise to four years by 2008 (most biotech and electrical patents already take four years to issue).
The system favors the [patent applicant] from the start. A patent examination is basically a secret proceeding. The patent is granted, unless the examiner can show that a patent should not be granted. An inexperienced examiner with a newly minted bachelor's degree in science is often pitted against a high-paid patent attorney. And if you are still refused a patent, you can file a continuation application, restarting the process and maybe getting an easier examiner. Counting continuations, 85 percent to 95 percent of applications in the U.S. are eventually granted--a much higher rate than in Europe or Japan.
Patentees and their lawyers are now addicted to a high-volume stream of easy-to-get patents.
I am reminded of the story during the Vietnam War when the White House sent two analysts, one a hawk and one a dove, on a fact-finding mission. Their reports diverged so sharply that the recipient asked, "Are you sure you two visited the same country?"
Finally, there is non-lawyer Paul Graham, who makes much more sense to me than either of the lawyers.
A lot of companies (Microsoft, for example) have been granted large numbers of preposterously over-broad patents, but they keep them mainly for defensive purposes. Like nuclear weapons, the main role of big companies' patent portfolios is to threaten anyone who attacks them with a counter-suit.