4 ms·
I think its possible to have a good software patent. However in my quite long experience in the software industry, I quite literally have never seen a software
by zmmmmm 7y ago
I think its possible to have a good software patent. However in my quite long experience in the software industry, I quite literally have never seen a software idea that was actually patented and which convincely passed the "skilled in the art" novelty test. Nearly everything I have seen would be logically thought of by someone even moderately skilled in the art if presented with the same requirements. So I think the conclusion is, even if there are a few "good apples" if the barrel is full of 98% "bad apples", you really should throw out the whole barrel.
- GoblinSlayer 7y agoMaybe apply some simple formal criterion, like if function is implemented in under 100kloc, then it's trivial and invalidates the patent.
- noahadavis 7y agoLOC requirements are easy to abuse and LOC != amount of originality
- greyfade 7y agoI think it's fair to say that LoC is perhaps the least relevant metric it is possible to imagine. Languages differ in expressive power in ways that effectively mean that 200kLoC in one language is trivial to implement in a few thousand lines in another. Even within one language, two implementations of the same effective feature can be devised with similarly large differences in line count, if one prioritizes brevity over efficiency. A simple criterion like this would be laughably useless and abused to absurd degrees.
- GoblinSlayer 7y agoI don't mean that 200kloc implementation proves non-triviality, I mean 100kloc implementation proves triviality. Also efficiency is achieved by implementation of specific algorithm, not by code formatting.