Xref: utzoo gnu.misc.discuss:3167 comp.misc:12623 comp.dcom.modems:9844 Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!news.larc.nasa.gov!ipsun.larc.nasa.gov!jcburt From: jcburt@ipsun.larc.nasa.gov (John Burton) Newsgroups: gnu.misc.discuss,comp.misc,comp.dcom.modems Subject: Re: hayes lawsuit Message-ID: <1991May20.135423.14522@news.larc.nasa.gov> Date: 20 May 91 13:54:23 GMT References: <14833@ulysses.att.com> Sender: news@news.larc.nasa.gov (USENET Network News) Distribution: usa Organization: NASA Langley Research Center, Hampton, VA USA Lines: 123 In article <14833@ulysses.att.com> smb@ulysses.att.com (Steven Bellovin) writes: >The fundamental question we have to answer is this: if software >patents are to be disallowed, what alternative mechanism would you >propose to encourage people to publish new algorithms? (I'd >rather argue the question using RSA as an example; its originality, >novelty, and utility are beyond question. The Hayes patent may >not qualify as new -- but the very difficulty of resolving that >issue makes it a bad hook upon which to hang the policy question.) > I think one problem with your argument/question is it presumes that the primary reason for "publishing a new algorithm" is for monetary gains. This is not always the case. As far as "publishing new algorithms", people have been doing that for years WITHOUT using patents. Its known as publication in a professional journal, which are generally protected under COPYRIGHT laws. If you want new algorithms, take a look at "Communications of the ACM"...no patents there. If you really take a look into who supports patents and who opposes them, you'd find that the business managers & lawyers (the ones who handle the MONEY side of things) are for software patents, the programmers & computer scientist (the ones who work with the computers) are against software patents. This doesn't always hold, but to a great extent it does. Tell me if this sounds crazy...I want to get a patent on the following... 1. Take two eggs, a cup of flower, two tablespoons of butter, a quarter cup of sugar, and a quarter cup of milk. 2. Mix it all together in a medium size bowl. 3. Put the mixture in a flat pan and bake for 20 minutes. 4. (heres the unique part) Cut it in half, and bake each half for another 20 minutes. This is an alogrithm...a series of instructions producing a desired result from a given starting point. Of course if I tried to get a patent on this, I would be laughed out of the patent office. Yet if this had been a computer program, I probably could have gotten a patent. The difference? the computer program could make someone a fortune, the recipe probably couldn't. In mathematics, algorithms are used all the time, with new ones being published almost continually. Can you imaging what would happen to ALL the scienes (and from there ALL technology and ALL society which is based on that technology) if patents were used to "encourage mathematicians to publish their algorithms"??? Lets see, since just about everything IN mathmatics IS an algorithm...hmmm..."well, I have a patent on Linear Interpolation, so everyone that uses Linear Interpolation will have to pay me a license fee..." "I have a patent on floating point multiplication...guess every computer manufacturer who makes a floating point coprocessor will have to pay me a license fee...boy, will I make MONEY!!!". If patents on algorithms had been used by mathematicians starting in the 1800's, science would still be back in the 1800's because instead of going forward, and doing basic research, the majority of manpower would be wasted trying to "reinvent the wheel" to get around the alogorithm patents. In the field of computer programming, if software patents become the rule, as opposed to the exception, all progress would come to grinding halt in a few years...more and more manpower would be applied to working around patents, and working around the patents on the work arounds, and working around....(you get the picture). So basically, I hope you are happy with the current state of computer software, because as software patents become more and more common, less and less progress will be made. Progress is adding another floor to a building made by others. If you had to go back and lay a new foundation each time, little progress would be made. The software patent does serve several purposes. It DOES provide jobs for patent lawyers. It DOES protect a company that is not able to compete technically. It DOES provide jobs for programming drones, whose only job is to keep "reinventing the wheel" in an effort to keep ahead of the patents from other companies. >I'm perfectly serious. In an industry where fortunes are being >made, and where many vendors have gone overboard on copy protection >because of piracy, how are we to encourage publication, if not >by patents? The issue of abuse, either statutory (i.e., a 17 >year period, which is too long for software) or in the Patent Office >(granting dubious patents due to lack of expertise) are, I would >say, separable questions. I don't know, seems to me, companies like Borland are doing fairly well without resorting to patents and litigation. There is a MAJOR difference between protecting a software package and patenting algorithms. A software package consists of possibly hundreds of individual alogrithms. Patenting the individual algorithms stifles progress (as I discussed above) and does not protect from piracy. Many companies are finding that copy protection isn't the answer since many users simply won't buy a package that is copy protected. You are correct, there is a problem with illegally copying software. Its different from illegally copying manuscripts (protected under copyright laws) in magnitude (i.e. much easier to electronically copy a huge program, as opposed to photocopying a huge manuscript), but perhaps copyright protection of the entire package would be a better starting point than patent protection of the individual algorithm. If a company profits from copying and redistributing anothers product, they should be legally prosecuted. If a company A creates a product that performs the same operations and functions as company B's product (but is NOT copied from company B), and sells it for less, then that should be considered fair competition. What *should* happen is that company B improves its product, either by cutting its price, making it faster or smaller or adding features so it would be more competitive with company A' product. What *does* happen is that company B sues company A. This results in added legal cost for both companies (creates more jobs for lawyers) and stifled competition. You ask why American companies can't compete in the world marketplace? They spend too much time an money in litigation and not enough in product research and development. This is not just true in the computer industry, but throughout the American economy. sorry, enough rambling...software patents are NOT the answer, despite what the lawyers say... John +--------------------------------------------------------------------+ | John Burton | | G & A Technical Software | | jcburt@gatsibm.larc.nasa.gov | | jcburt@cs.wm.edu | | | | Disclaimer: Hey, what can I say...These are *my* views, not those | | of anyone else, be they employer, school, or government| +--------------------------------------------------------------------+