3 ms·

I agree: don't make the concept/idea patentable, only the specific implementation. The specific light bulb; the specific telegraphy system. OTOH, I kinda like

by 13ren 18y ago

I agree: don't make the concept/idea patentable, only the specific implementation. The specific light bulb; the specific telegraphy system. OTOH, I kinda like the idea of Segway owning the concept - it was a cool idea, and a lot of work to pull it off. But I don't like the world that owning concepts so generally would create. I think it's good to encourage many people to search for ways to do it, and they get to own what they find. If someone finds the best way, they win big. I think that's great. I find it interesting that Morse's original patent seemed pretty much along the lines you say - but then, several years later, he tried to amend it to cover any way of accomplishing telegraphy. The courts upheld the original, but not the amendment (if memory serves). The amendment seems to be the style of patenting software today. I see no reason, in principle, why patent law can not operate in the same way in the software world as it does for mechanical systems - despite software being more abstract.

noonespecial 18y ago
I think that software patents have failed us because code is still magic to the USPO. One computer that, for example, calculates large prime numbers looks exactly like another that does the same. The patent system considers the one a copy of the other, It cannot understand that there are radically different ways to calculate large primes. In short they actually don't understand software at all and try to treat computers like gear filled machines.
- 13ren 18y ago
  You're right. In a sense, that's also true of comp sci academics and professionals. The field isn't a disciplined field (unlike mechanics). I think it's partly immaturity (mechanics is older); partly massive opportunity (everyone just reinvents everything); and partly the mathematical nature of comp sci... In my research, I've seen mathematical models reinvented with different technologies a decade later: e.g. first SGML and OODBMS; now XML and Java. The later writers are unaware of the earlier writers. There's very little knowledge of what's been published. It's hard to find, and hard to recognize because of different terminology. It can be very hard to recognize in a deeper sense, whether two domains (and algorithms operating on them) are actually equivalent (e.g. languages recognized by automata vs. languages generated by regular expressions). BTW: Ken Thompson patented his method for searching with regular expressions (expired 1988). http://www.google.com/patents?id=Cx9xAAAAEBAJ&printsec=abstract&zoom=4&dq=3568156#PPA5,M1 http://www.google.com/patents?id=Cx9xAAAAEBAJ&printsec=a... As far as understanding the fundamentals of computer science, I think we are very poor at it - much of the mathematics is about very simple and basic stuff (though the mathematics itself is usually not so simple...). Mechanics can also be described by mathematics - arguably, is mathematics in a sense. I don't think that being mathematical is a killer for computer science patents.

by 13ren 18y ago

noonespecial 18y ago

I think that software patents have failed us because code is still magic to the USPO. One computer that, for example, calculates large prime numbers looks exactly like another that does the same. The patent system considers the one a copy of the other, It cannot understand that there are radically different ways to calculate large primes. In short they actually don't understand software at all and try to treat computers like gear filled machines.

13ren 18y ago
You're right. In a sense, that's also true of comp sci academics and professionals. The field isn't a disciplined field (unlike mechanics). I think it's partly immaturity (mechanics is older); partly massive opportunity (everyone just reinvents everything); and partly the mathematical nature of comp sci... In my research, I've seen mathematical models reinvented with different technologies a decade later: e.g. first SGML and OODBMS; now XML and Java. The later writers are unaware of the earlier writers. There's very little knowledge of what's been published. It's hard to find, and hard to recognize because of different terminology. It can be very hard to recognize in a deeper sense, whether two domains (and algorithms operating on them) are actually equivalent (e.g. languages recognized by automata vs. languages generated by regular expressions). BTW: Ken Thompson patented his method for searching with regular expressions (expired 1988). http://www.google.com/patents?id=Cx9xAAAAEBAJ&printsec=abstract&zoom=4&dq=3568156#PPA5,M1 http://www.google.com/patents?id=Cx9xAAAAEBAJ&printsec=a... As far as understanding the fundamentals of computer science, I think we are very poor at it - much of the mathematics is about very simple and basic stuff (though the mathematics itself is usually not so simple...). Mechanics can also be described by mathematics - arguably, is mathematics in a sense. I don't think that being mathematical is a killer for computer science patents.