Hal Finney [hal@rain.org] wrote:
>In cryptography, the researcher sometimes called the "patron saint"
>of cypherpunks, David Chaum, has a number of patents on technologies
>which can greatly enhance freedom and privacy. It appears that the
>hope of patent protection has been a significant motivation for him.
>This is what you risk losing if you eliminate patents.
Mark, mark@unicorn.com, replies:
> Citing Chaum as an argument for patents seems bizarre to me. Last time I
> hung out with a bunch of cypherpunks, most times Chaum's name was mentioned
> were in comments like: 'Chaum's patents are a major cause of the lack of
> privacy on the Net', 'Do you think Chaum's being paid to keep anonymous
> ecash off the market?', 'How can we work around Chaum's patents and set up
> our own anonymous ecash system?' and 'When do Chaum's patents expire so we
> can have anonymous ecash at last?'
Much of the unhappiness with Chaum has not been with the patents themselves, but with his uncooperative licensing terms. The general perception is that Chaum has not proved to be a competent businessman, and that as a result he has badly mismanaged his patent portfolio. In fact, his company, DigiCash, has recently filed for bankruptcy, and there is a general belief that whomever ends up owning his patents in the end will make them much more widely available.
> The ecash blinding patent isn't as obvious as an XOR cursor, but it's not
> far from it. If Chaum hadn't patented it someone else would have discovered
> it by now and it would probably be widely used in anonymous systems; but
> because Chaum did patent it the technology is languishing and won't see
> widespread use until the patent expires. This seems to be a major argument
> *against* patents, not for them.
It's true that there is a risk that the patent holder will not behave rationally so as to maximize profits from his patent. As in other types of free-market irrational behavior, sooner or later the market corrects, as we are seeing now with the DigiCash bankruptcy. Unfortunately, it can easily take a few years, but that's just the inherent time scale of human affairs.
> On the more general front, the main problem with software and computer
> hardware patents is that twenty-year monopolies are absurd in an industry
> where the technology changes every year or two; a twenty-year monopoly is
> effectively infinite, because the patented technology will almost certainly
> be obsolete by the time the patent expires. This is not true in, say, auto
> manufacture, where the basic technology has changed little in the last
> century.
I disagree about the patented technology becoming obsolete. The patents people complain most about are fundamental, and even new technologies will build on them. For example, the patent on public key cryptography techniques expired last year, but public key crypto will be used for decades, at least. The patent on RSA expires in 2000, but that technology will certainly still be used. In fact, for both of these cases you can argue that the patents were too early, with large-scale use of these technologies coming only in the last few years of the patent.
Chaum's blinding-signature patent, which covers the basic idea of ecash, expires in 2005. The technology will surely continue to be useful at that time.
Obviously computer technology changes a great deal over a 20 year period. But software algorithms do not have a similar rate of change. In fact, if algorithms became obsolete every two years the way computers do, people probably wouldn't care about such patents. The reason they do care is exactly because algorithms patented 20 years ago are still useful today.
The 20 year period is based on the time it takes to set up a business, get it capitalized, get product into production, amortize startup costs, and see a useful return from the investment in the patented technology. There may also be a time delay before the whole process can begin, as you wait for the patent itself to become practical. This would be the case for cutting-edge research, like the public key cryptography patent.
Here you might make a stronger argument that software is different from other patented technologies, because software companies have lower startup costs and perhaps can get into production more quickly. It could be the case that a software patent begins paying back after only a couple of years, rather than a decade or so.
On the other hand, I read in Wired yesterday that the software business is obsolete. :-) Nobody is making any money introducing new software products (except for games). Everything is on the web. All the big, successful startups are giving away their software. That's the only way to build volume these days.
Of course, we take such broad statements with a large grain of salt. But it points out that it is not often the case that a software patent, by itself, turns into a saleable product. It may be just one ingredient in a larger system; the final end product may not be a piece of software at all, but a device with some software in it. In those cases, the payback time is going to be more similar to traditional business models.
It does seem that everything is moving faster today; in any business, software or otherwise, you can get started, make a buck, and move on, much more quickly than decades ago. It could be that the 20 year period could be reduced to 15 or even 10 years and still provide a reasonable return on investment - although, as mentioned above, this would have made the PKC and RSA patents virtually valueless.
Hal