DB Research has issued a paper that claims software patents block innovation:
A growing number of R&D-intensive businesses realises that licencing out their IP (intellectual property) can constitute a substantial share or their revenues, which in turn encourages innovation efforts. Bearing this in mind, one could be tempted to consider ever stricter IP protection regimes to provide ever more stimuli for innovation.
This conclusion is wrong, however. A prime example is patents on software, which might at first sight be seen as a logical expansion of the classic technology patent. But creating software differs markedly from creating machinery and the like: MIT researchers Bessen and Maskin argue that innovation in software is both strongly sequential (one invention building on a previous one) and complementary (thriving on parallel approaches to the same problem), far more so than in other technology fields. In fact, they found empirical evidence that software patenting substitutes R&D activity, rather than encouraging it, and conclude: “For industries like software or computer, there is actually good reason to believe that imitation becomes a spur to innovation, while strong patents become an impediment” 2. In accordance with other academics, they strongly favour copyright over patent protection for software. Copyrighting provides both adequate leeway for sequential innovation and enough protection for marketable software products.
Update: Siegfried Hirsch posted a german translation of this passage.