andrew@stc.UUCP (Andrew Macpherson) (07/15/85)
In article <9028@ucbvax.ARPA> phr@ucbvax.ARPA (Paul Rubin) writes: >The following letter appeared in the July, 1985 issue of BYTE magazine. > ... The letter: > > ... Charles Kluepfel described an > implementation of the RSA Public Key algorithm and the BASIC code required. > Unfortunately, he did not reference that this RSA Public Key Cryptosystem > was patented by the Massachusetts Institute of Technology in 1983 > (U.S. Patent 4,405,829). The worldwide exclusive license to this patent > was then purchased from MIT by RSA Security Inc., a company founded by > the inventors of the RSA algorithm to develop this technology. > ... Rather, the purpose is to make you aware of our > patent position and ask for your help in educating your readership as > to its existence. Based on Mr. Kluepfel's article, more people are > going to start expending money and effort developing RSA-based software > for commercial purposes. Regrettably, their effort will be wasted > unless they obtain a sublicense from us. This seems strange. I was under the impression that: 1 algorithms cannot be patented, hence the use of copyright and/or trade secret law to protect software. 2 software independently developed to perform a given function is the property (copyright etc.) of the developer to use as he sees fit, and it is only that based on someone else's *CODE* which is open to legal attack. Since I don't claim to be a Legal expert, much less an American Legal expert (US Patent after all :-) I'm cross-posting to net.legal, with followup to net.legal only. I'ld be interested also in comments from anyone qualified in the legal aspects in europe. -- Regards, Andrew Macpherson. <andrew@stc.UUCP> {creed, datlog, idec, iclbra, root44, stl, ukc}!stc!andrew
hal@cornell.UUCP (Hal Perkins) (07/15/85)
from net.crypt: > Unfortunately, he did not reference that this RSA Public Key Cryptosystem > was patented by the Massachusetts Institute of Technology in 1983 > (U.S. Patent 4,405,829). The worldwide exclusive license to this patent > was then purchased from MIT by RSA Security Inc., a company founded by > the inventors of the RSA algorithm to develop this technology. > > Because the RSA algorithm has been published in academic journals, most > people assume that it is in the public domain, similar to the DES > algorithm. Unfortunately, some people have developed software and > other products based on the RSA algorithm without researching this > point. Nevertheless, the patent exists and, in the opinion of our > corporate attorneys, will be easily defended. As RSA Security Inc. > paid a great deal of money for the exclusive patent rights, we plan > to actively police the commercial use of the RSA algorithm. Er, say what? I'm no lawyer, but from my reading of general articles on patent law, an algorithm is one of the things that specifically CAN'T be patented. One can patent a gadget (that's how the Unibus was patented), but one can't patent an idea unless it's "reduced to practice", i.e., implemented. That would seem to mean that one could patent a box that encrypted data using the RSA algorithm, but one couldn't patent the algorithm itself. What's going on here? (I've aimed this at net.legal also and I think followups will go there since that seems to be the appropriate place for this.) Hal Perkins UUCP: {decvax|vax135|...}!cornell!hal Cornell Computer Science ARPA: hal@cornell BITNET: hal@crnlcs
smb@ulysses.UUCP (Steven Bellovin) (07/16/85)
> I'm no lawyer, but from my reading of general articles on patent law, > an algorithm is one of the things that specifically CAN'T be patented. > One can patent a gadget (that's how the Unibus was patented), but one > can't patent an idea unless it's "reduced to practice", i.e., implemented. > That would seem to mean that one could patent a box that encrypted data > using the RSA algorithm, but one couldn't patent the algorithm itself. > What's going on here? The patentability of algorithms is still an unsettled question. At least, that was the impression I got from the USENIX tutorial on intellectual property in Portland.
b-davis@utah-cs.UUCP (Brad Davis) (07/16/85)
In article <3154@cornell.UUCP> hal@gvax.UUCP (Hal Perkins) writes: >from net.crypt: >> Unfortunately, he did not reference that this RSA Public Key Cryptosystem >> was patented by the Massachusetts Institute of Technology in 1983 >> (U.S. Patent 4,405,829). The worldwide exclusive license to this patent >> was then purchased from MIT by RSA Security Inc., a company founded by >> the inventors of the RSA algorithm to develop this technology. >> >I'm no lawyer, but from my reading of general articles on patent law, >an algorithm is one of the things that specifically CAN'T be patented. I'm no lawyer either but a patent attorney told me that algorithms are not patentable. In fact only firmware that is tightly bound to patentable hardware will even be considered. I think that RSA PKC is a hardware device with some scare tactics to stifle competition. -- Brad Davis {ihnp4, decvax, seismo}!utah-cs!b-davis b-davis@utah-cs.ARPA
karn@petrus.UUCP (Phil R. Karn) (07/17/85)
Along the same lines, IBM holds patents on DES. It is claimed that they will grant royalty-free licenses to anyone wishing to manufacture chips conforming to the standard. A few thoughts come to mind: 1. To formally adhere to DES (i.e., to gain a certification from NBS), you have to implement the algorithm in hardware on a special-purpose chip. General purpose computer software implementations are not certifiable. I've never understood this requirement, except to note that a hardware implementation seems to be related to the test for patentability. 2. IBM is not widely known to cheerfully give away things for free. In fact, this is so out of character for them that one almost wonders if they have an ulterior motive to promote widespread usage of DES (I think you can guess what it might be.) Or, more charitably, perhaps they realize their patent on the "algorithm" itself could never be defended anyway, so they hope to gain some public good will out of it. Naturally, the above are strictly my personal opinions. Phil
henry@utzoo.UUCP (Henry Spencer) (07/18/85)
Doubtless the legal eagles will correct me if I blow this, but into the fray anyway... > 1 algorithms cannot be patented, hence the use of copyright > and/or trade secret law to protect software. The theoretical intent of patent law is to protect inventions, i.e. new and useful ways of doing things. At least some algorithms clearly ought to qualify. In reality, the situation is rather unclear. Scientific discoveries, i.e. laws of nature, definitely are not supposed to be patentable, and there is a problem of definition here: just what exactly are algorithms? In practice, you can probably patent an algorithm which is expressed as a hardware implementation (it's been done), and you might be able to use that patent to sue people who implement the same thing in software. I would guess that this is the situation for the RSA encryption scheme. One reason why copyright and/or trade secret law are the primary means for protection of software is that most pieces of software are not innovative enough to qualify as inventions. They are routine exercises of technical skill, rather than striking new inspirations. Another reason is that the murky situation regarding patents on algorithms has scared people off; trade secrets don't have that kind of legal confusion surrounding them. A third reason is that our patent system has been allowed to deteriorate to the point where it no longer provides much protection to inventors, and trade secrets are once again on the rise as the only effective protection method (patents were originally devised to make it possible to protect inventions *without* keeping them secret). > 2 software independently developed to perform a given function > is the property (copyright etc.) of the developer to use as > he sees fit, and it is only that based on someone else's > *CODE* which is open to legal attack. "But I did it myself, borrowing only his idea" is insufficient; it is the idea itself that is protected by patents, not the details of its expression. (Copyrights, on the other hand, protect form of expression rather than underlying concepts.) In fact, completely independent invention of the basic idea is not a defence against patent infringement either. Once the inventor has a patent on his invention, it is *his* *property* until the patent expires. Exclusively his. You are thinking of things like the Unix licence, which are based on trade-secret law; independent invention *is* a defence there. -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,linus,decvax}!utzoo!henry
inc@fluke.UUCP (Gary Benson) (07/19/85)
>> I'm no lawyer, but from my reading of general articles on patent law, >> an algorithm is one of the things that specifically CAN'T be patented. > I'm no lawyer either but a patent attorney told me that algorithms are > not patentable. In fact only firmware that is tightly bound to patentable > hardware will even be considered. Well I'm no lawyer either, and I soemtimes wonder if any of them read this newsgroup! It might be interesting to read what an attorney (as opposed to a friend of an attorney) has to say on legal issues. Now I won't suggest that your friend probably got his(her) law degree out of a Cracker Jack box, but in fact algorithms *are* patentable. It is only recently that this has come about, so not many have been done. Now then, I'm no doctor, but I've got a theory about hernias... *** REPLACE THIS LINE WITH YOUR MESSAGE *** -- Gary Benson * John Fluke Mfg. Co. * PO Box C9090 * Everett WA * 98206 MS/232-E = = {allegra} {uw-beaver} !fluke!inc = = (206)356-5367 _-_-_-_-_-_-_-_-ascii is our god and unix is his profit-_-_-_-_-_-_-_-_-_-_-_
magik@wlcrjs.UUCP (Ben Liberman) (07/19/85)
[] The heading on the patent is: CRYPTOGRAPHIC COMMUNICATIONS SYSTEM AND METHOD I pulled this from the Official Gazette which is the patent offices listing of patent abstracts. There isn't enough detail here to tell much but I'd place some emphasis on the AND in the title and assume it's a pachage deal. -- ----------------------------------------- Ben Liberman {ihnp4|ihldt}!wlcrjs!magik
magik@wlcrjs.UUCP (Ben Liberman) (07/19/85)
[] The heading on the patent is: CRYPTOGRAPHIC COMMUNICATIONS SYSTEM AND METHOD I pulled this from the Official Gazette which is the patent offices listing of patent abstracts. There isn't enough detail here to tell much but I'd place some emphasis on the AND in the title and assume it's a pachage deal. -- -- ----------------------------------------- Ben Liberman {ihnp4|ihldt}!wlcrjs!magik
sean@ukma.UUCP (Sean Casey) (07/23/85)
In article <405@petrus.UUCP> karn@petrus.UUCP (Phil R. Karn) writes: >2. IBM is not widely known to cheerfully give away things for free. In >fact, this is so out of character for them that one almost wonders if they >have an ulterior motive to promote widespread usage of DES (I think you can >guess what it might be.)... You bet. The NSA really boondoggled IBM about DES and it's implementation. Could it be that once upon a time it was so good that the NSA was forced to cripple it? -- - Sean Casey UUCP: sean@ukma.UUCP or - Department of Mathematics {cbosgd,anlams,hasmed}!ukma!sean - University of Kentucky ARPA: ukma!sean@ANL-MCS.ARPA
peter@baylor.UUCP (Peter da Silva) (07/24/85)
> One reason why copyright and/or trade secret law are the primary means > for protection of software is that most pieces of software are not > innovative enough to qualify as inventions. They are routine exercises > of technical skill, rather than striking new inspirations. Another > reason is that the murky situation regarding patents on algorithms has I've got news for you... LOTS of patented inventions are just routine excersises of technical skill. Just look at some of the junk that has patents on it... Software just seems to confuse patent attornys enough that it's gotten special status. -- -- Peter da Silva (the mad Australian) -- UUCP: ...!shell!neuro1!{hyd-ptd,baylor,datafac}!peter -- ARPA: baylor.peter@RICE.ARPA -- MCI: PDASILVA; CIS: 70216,1076; DELPHI: PJDASILVA --
bobh@pedsgd.UUCP (Bob Halloran) (07/25/85)
In article <1982@ukma.UUCP> sean@ukma.UUCP (Sean Casey) writes: >In article <405@petrus.UUCP> karn@petrus.UUCP (Phil R. Karn) writes: >>2. IBM is not widely known to cheerfully give away things for free. In >>fact, this is so out of character for them that one almost wonders if they >>have an ulterior motive to promote widespread usage of DES (I think you can >>guess what it might be.)... > >You bet. The NSA really boondoggled IBM about DES and it's implementation. >Could it be that once upon a time it was so good that the NSA was forced to >cripple it? > It is widely believed that the original IBM proposal involved a key size of more than 128 bits, and that the No Such Agency promptly classified the research and imposed the existing size in order to facilitate their domestic monitoring of traffic. Bob Halloran Sr MTS, Perkin-Elmer DSG ============================================================================= UUCP: {ihnp4, decvax, ucbvax}!vax135!petsd!pedsgd!bobh USPS: 106 Apple St M/S 305, Tinton Falls NJ 07724 DDD: (201) 758-7000 Disclaimer: My opinions are mine alone. Quote: "No matter where you go, there you are" - B. Banzai
jim@randvax.UUCP (Jim Gillogly) (07/25/85)
I read the letter claiming a patent on the RSA algorithm and passed it to Stockton Gaines, who knows about things like this. His first impression was the one given by several people on the net, that you can't patent an algorithm or an idea. However, since then he ordered and got a copy of the patent and has changed his mind. Apparently one kind of legally acceptable patent is the "use" patent, which lays claim to certain uses of something. The example he gave is an inventor who discovers that a chemical can be used for something unexpected. For example, if you discover that aspartame will behave as a room-temperature superconductor (different from its use as a sweetener), you can patent any use of aspartame as a conductor. These people have patented the use of the RSA algorithm in encryption. Apparently if you want to use it as a good but slow random number generator you're still in the clear ... as long as the random numbers aren't used to encrypt plaintext! I'd be interested in reactions to this theory... -- Jim Gillogly {decvax, vortex}!randvax!jim jim@rand-unix.arpa
dee@cca.UUCP (Donald Eastlake) (07/28/85)
The reason certified implementations of DES have to be in hardware is because it is much more tamper resistant than software. It was part of the deal that IBM had to agree to give royalty free licenses to DES for it to be designated a standard. IBM uses DES on all of its internal communications lines. -- +1 617-492-8860 Donald E. Eastlake, III ARPA: dee@CCA-UNIX usenet: {decvax,linus}!cca!dee
bwm@ccice1.UUCP (Bradford W. Miller) (07/30/85)
In article <301@baylor.UUCP> peter@baylor.UUCP (Peter da Silva) writes: >> One reason why copyright and/or trade secret law are the primary means >> for protection of software is that most pieces of software are not >> innovative enough to qualify as inventions. They are routine exercises >> of technical skill, rather than striking new inspirations. Another >> reason is that the murky situation regarding patents on algorithms has > >I've got news for you... LOTS of patented inventions are just routine >excersises of technical skill. Just look at some of the junk that has patents >on it... Software just seems to confuse patent attornys enough that it's >gotten special status. This is a good point. In fact, most patents do not in fact hold up in court. It's mainly a scare tactic. The patent office certainly does not know what is a 'routine exercise' so having a patent deters those who are afraid of 'breaking' a patent - though it wouldn't hold up. Brad Miller -- ..[cbrma, ccivax, ccicpg, rayssd, ritcv, rlgvax, rochester]!ccice5!ccice1!bwm
tim@callan.UUCP (Tim Smith) (08/27/85)
> I'm no lawyer, but from my reading of general articles on patent law, > an algorithm is one of the things that specifically CAN'T be patented. I just recently found a book at the Caltech bookstore called "Software Law, a Primer", and bought it. It was quite fascinating. One of the sample cases given was some sort of control system for a chemical plant that used a specific equation to control something or other. This was patentable. The patent does not cover all uses of that equation - just the use of it control that specific part of a chemical plant doing what that specific chemical plant was doing ( I don't have the book with me, so I am being a bit vague here. Sorry ). Although an algorithm can't be patented ( I think ), using a specific algorithm to accomplish a specific task can. This is probably what will make the RSA system patentable. DISCALIMER: I am not a lawyer, or even a particularly well read non-lawyer when it comes to this sort of thing. -- Tim Smith ihnp4!{cithep,wlbr!callan}!tim
mpr@mb2c.UUCP (Mark Reina) (08/29/85)
> > I'm no lawyer, but from my reading of general articles on patent law, > > an algorithm is one of the things that specifically CAN'T be patented. > > I just recently found a book at the Caltech bookstore called "Software Law, > a Primer", and bought it. It was quite fascinating. > One of the sample cases given was some sort of control system for a chemical > plant that used a specific equation to control something or other. This was > patentable. The patent does not cover all uses of that equation - just the > use of it control that specific part of a chemical plant doing what that > specific chemical plant was doing ( I don't have the book with me, so I am > being a bit vague here. Sorry ). > I am quite sure that an algorithm can not be patented. However, the process can be patented. For instance, when the telephone was patented it made perfect use of applied science. The process of turning a "voice" into an electrical pulse and converting it back again. The patent did not cover that algorithm. But the tangible product making the conversion was patented. This may be one of the most valuable patents ever issued under the U. S. Patent Office. Mark Reina
smb@ulysses.UUCP (Steven Bellovin) (08/29/85)
> > I'm no lawyer, but from my reading of general articles on patent law, > > an algorithm is one of the things that specifically CAN'T be patented. > > .... > > Although an algorithm can't be patented ( I think ), using a specific > algorithm to accomplish a specific task can. This is probably what > will make the RSA system patentable. That is indeed what they did. The RSA patent is for a communications system that uses the equations to accomplish assorted stuff, like public keys, signatures, etc. If you can think up new uses for their equations that are neither covered by the patent nor an "obvious" extension, you're home free.
doc@cxsea.UUCP (Documentation ) (08/31/85)
> > I'm no lawyer, but from my reading of general articles on patent law, > > an algorithm is one of the things that specifically CAN'T be patented. > > I just recently found a book at the Caltech bookstore called "Software Law, > a Primer", and bought it. It was quite fascinating. > > One of the sample cases given was some sort of control system for a chemical > plant that used a specific equation to control something or other. This was > patentable. The patent does not cover all uses of that equation - just the > use of it control that specific part of a chemical plant doing what that > specific chemical plant was doing ( I don't have the book with me, so I am > being a bit vague here. Sorry ). > > Although an algorithm can't be patented ( I think ), using a specific > algorithm to accomplish a specific task can. This is probably what > will make the RSA system patentable. > > DISCALIMER: I am not a lawyer, or even a particularly well read non-lawyer > when it comes to this sort of thing. > This is one of those odd-ball legal areas that causes more trouble than it has any right to. About 15 years ago, someone sought to patent a simple program for converting binary numbers to decimal ones. The Supreme Court eventually decided that this was not patentable, because the program used iterations of a simple formula to do the conversion. The court concluded that this "algorithm", like most mathematics, is a law of nature, which is clearly non-patentable. The problem has been that people tend to confuse the court's use of the word "algorithm" with the software developer's use of the word "algorithm", which are in fact different. You might consider a re-entrant C function to be an "algorithm", but what does that have to do with a mathematical formula? I suppose you could create a mathematical model of the function, but you could also model anything else mathematically, such as a chemical process or a mechanical device, both of which are normally patentable. But the "program is an algorithm is a mathematical formula" idea just wouldn't go away. So the courts and the patent office were stuck for awhile with this notion of "program" as synonymous with "law of nature", and said "no patent". Well, since then, things have changed. The courts have ben convinced that software is not a simple law of nature, anymore than any other mechanical, electrical or chemical contrivance is. The Patent Office has routinely accepted software patent applications since 1982, which presumably cover all kinds of program "algorithms". So, yes, you can patent a software algorithm, so long as it is something more than a simple formula or equation, assuming it meets the statutory requirements for patentability (non-obviousness, novel, etc.)
edhall@randvax.UUCP (Ed Hall) (09/04/85)
I guess it depends upon what you call an ``algorithm''. I understand that UNIX's set-user-ID mechanism was one of the first (if not THE first) software patents. It has nothing to do with particular hardware or servicing a particular application--I would assume that a computer implemented with bistable algae built solely for performing a simulation of water flow is still covered by this patent, should it require such a file-security mechanism. -Ed Hall decvax!randvax!edhall