A Patent for Trouble

At Techsummit 2005 one of the big topics was software patents. Pretty much everyone knows there are problems with them; for example, a British firm recently tried to patent hyperlinks (I believe that one was rejected by a court), and many dubious patents have been approved by the U.S. Patent and Trademark Office (USPTO), the most famous of which are probably Amazon's One Click Shopping patent and one for online credit card authorization. Such patents promote monoply and thus monoculture, which makes software, computers, the Internet, and the economy brittle and at risk. You can fight such patents after the fact, as EFF is doing, or perhaps more radical solutions are called for.

I thought I might have to summarize the whole lengthy issue, but as it happens, the Economist has just published a lengthy survey in several articles. A really brief summary might include that patents were invented to promote invention and thus to promote the economy by giving inventors temporary limited rights to their inventions so they would have economic incentive to invent them. Software is a much newer invention than patents, and software patents are much newer than software.

The first software patent I ever heard of was snuck through by including diagrams and description as if it were a physical machine: this was Dennis Ritchie's patent no. 4,135,240 of 16 January 1979, "protection of data file contents" for the Unix setuid bit. Ironically, much early development of Unix was funded by its first paying application, which was a text formatting system that was used for formatting patents.

Software patents did not become popular in the U.S. until the Supreme Court ruled in Diamond v. Diehr in 1981 that software was patentable and Congress appointed a court in 1982 to deal with patents.

Software is all about processes, processes which can be implemented in numerous malleable ways, probably more malleable than hardware implementations, and software involves numerous subdiscipilnes, too many to expect any patent examiner to learn.

But the big problem with software patents is that big businesses have taken to using them as weapons to keep smaller businesses out. Patents as weapons are nothing new. In the mid nineteenth century the two main manufacturers of sewing machines, Elias Howe and Isaac Singer, had to cross-license patents to each other so both would have sufficient permissions to actually manufacture sewing machines. This is called a patent pool or a patent thicket, and it serves to keep any of several vendors from suing each other, since each owns a patent on an essential piece of technology in their industry.

Patents as weapons to suppress competition are also not new. The Motion Picture Patents Company (the Trust) formed in 1908 to require license fees on Edison's patents from all film producers, distributors, or exhibitors. This legal monopoly eventually led to a bunch of pirate exhibitors fleeing to an orange grove on the west coast where they produced their own movies, and became known as Hollywood. They eventually legally overturned the Trust. Meanwhile, they had also invented the feature length motion picture while the Trust's members stuck to the old short format. Both innovation and economy had to work around a patent monopoly.

This last problem seems to be occuring again, this time with software patents. to quote Bill Gates:

If people had understood how patents would be granted when most of today's ideas were invented and had taken out patents, the industry would be at a complete standstill today. ... The solution is patenting as much as we can. A future startup with no patents of its own will be forced to pay whatever price the giants choose to impose. That price might be high. Established companies have an interest in excluding future competitors.
--Bill Gates, in an internal Microsoft memo of 1991, quoted by Fred Warshofsky in "The Patent Wars" of 1994

Ideally, the solution would be to tighten up patent approvals so that bad patents wouldn't get through, and then while there would no doubt still be griping about patents limiting competition, but the situation would be no worse than for any other kind of patents. Unfortunately, that solution is very hard to implement; there are no funds for additional or additionally trained patent examiners. There apparently have been attempts to mobilize changes in the patent process, but the other big users of patents, biotech and pharmacology companies, don't have this problem (their companies typically have only a few relevant patents to deal with, while software companies have sometimes thousands). Plus software companies and their sources of revenue, i.e., venture capitalists, are not terribly organized; for example a VC can always find a different sub-field to invest in if one has become constipated due to patents.

So an increasingly popular solution is to try to eliminate software patents entirely. The European Union (EU) theoretically doesn't have them, and India, along with many other countries, refuses to. It is rumored that some large companies deliberately file large numbers of dubious software patents in order to deliberately clog the system in hopes that it will fail. Others hope to organize patent pools or other means of coordinating intellectual property so as to work around the software patent system.

I will add that this whole controversy isn't about whether intellectual property (IP) is good or bad. Copyright is almost universally admired by the same people who want to abolish software patents, for example. Even the Free Software Foundation's Copyleft, or GNU Public License (GPL) is a sort of backhanded homage to intellectual property, attempting to protect it as a public good. Which is not to say that copyright doesn't have its own problems, but that's another subject, having much more to do with movies than with software.

It's software patents that are particularly broken, posing a risk of monopoly and thus monoculture, and the question is how do we fix them?

-jsq

Comments

[Alex Macfie]

You wrote...."a British firm recently tried to patent hyperlinks (I believe that one was rejected by a court)"

If you mean British Telecom's patent claim on hyperlinks, then you're not exactly right. What happened was... Some time in 1999 or 2000 (not really "recent" in internet terms), BT discovered it had an old US patent (dating from the mid-1970s) on hyperlinks, and decided it would enforce it against ISPs in the US. BT finally had its day in court in 2002, and lost decisively. Not because the patent was invalid, but because it was found not to cover the internet. BT's patent describes a hub-and-spokes type network with interlinked documents stored on the hub, which is quite obviously not how the internet works. See http://www.nysd.uscourts.gov/courtweb/pdf/D02NYSC/02-07733.PDF for the judgement against BT.

[james]

I saw that you have a page that discusses patent-related resources at http://riskman.typepad.com/perilocity/2005/10/a_patent_for_tr.html. I wanted to suggest adding www.freepatentsonline.com to the page. This web site has free PDF downloading (instead of having to page through TIFFs like at the US PTO). It is by far the best free patent searching site.