5 ms·

Software patents don't parallel hardware patents sufficiently closely. When you are granted a hardware patent, most everything a competent competitor needs to

by Hoff 16y ago

Software patents don't parallel hardware patents sufficiently closely. When you are granted a hardware patent, most everything a competent competitor needs to reproduce the invention is directly in the patent. With software (algorithm) patents, that's rather less the case. If you were to more closely parallel the hardware model and if you wanted to encourage the public good, then the grant of a software patent would escrow full source code, and (when the patent lapses) that same full source code is then available for use by others. And yes, I'm looking straight at the One Click patent here.

mseebach 16y ago
So why would source code for the "One Click" patent make any difference? It would take be all of ten minutes to make a perfectly functional implementation of "one click shopping" in an open source PHP cart which I could submit with the patent (or a day to build a simple cart from scratch if the open source-part is a problem). How would that make the patent less viable? Or are you arguing that only specific implementations could be patented? That, if we allow the premise of software patents in the first place, doesn't make any sense. Then we'd need measures to determine how much to change before it's not the same. I mean, to keep with the example of hardware patents, I'm sure a patent for a new crankshaft would also cover if it's 5% larger and made of a slightly different alloy. Re-implementing in an other language? What if it compiles to the same VM code? Then I'm getting the the business of generating and selling unique JVM languages. A new language for implementing the core of a competing algorithm? $1500, thank you.
- Vivtek 16y ago
  A patent for a new crankshaft might cover if it's 5% larger; it depends on how specific the claims are. There's a balance between too specific (thus useless) and too broad (thus not granted unless it's for freaking software); for machines this seems to be understandable, while for software the PTO just throws caution out the window, takes the money, and runs. The thing is, only a court decision will determine whether a patent covers a given use or not; I see a lot of documents written from both perspectives, and the real winners in the patent industry are lawyers (and, when all goes well and the case is sufficiently international, we translators get some nice money out of it, too).
- Hoff 16y ago
  Why? Because it (also) places indirect costs and additional onus onto the submitter of the patent; costs beyond the existing filing fees and legal fees. Because it also provides (some) tangible results for the community. (Now whether the covered source code is judged "good" or "bad" is another matter; there's certainly going to be some cases and some judgement required here around what's covered code, and whether or not the code is obfuscated, etc.) Because it means all of the patent-covered source code eventually gets exposed and licensed, which then gives an advantage to the community. Because even if your organization doesn't produce a product (and for whatever reason), you still have to produce functional source code. And because it means that most closed-source organizations probably won't choose to patent their software.
  - mseebach 16y ago
    In your post you referred prominently to the one-click patent. I fail to see how your proposal would have even slightly curtailed that patent. > Because it (also) places indirect costs and additional onus onto the submitter of the patent; A developer for a day, as I demonstrated. > Because it also provides (some) tangible results for the community. > Because it means all of the patent-covered source code eventually gets exposed and licensed An implementation of an obvious process isn't a tangible, or even particularly valuable, result, as any member of the community could produce it in very little time. > Because even if your organization doesn't produce a product (and for whatever reason), you still have to produce functional source code. No, you just have to throw a $500 project on e-lance. Or about what you pay the lawyer who's writing the patent for you every hour. > And because it means that most closed-source organizations probably won't choose to patent their software. No, at best it would keep the stuff that's actually patent-worthy from being patented (and thus "out there", inspiring others, which is the whole idea behind patents), while the obvious crap gets patented (what does Amazon have to lose from spending a day writing a simple PHP cart?)..
    - Hoff 16y ago
      This scheme makes patents part of adding to open-source, and also uses the comparative "allergy" that many commercial entities have to open source. The folks seeking patents will have to decide if they want the patent enough to open-source their covered source code (later). Various firms certainly won't want that. Those that do go for the patents will then serve to produce open-source code. But if you don't think it'll work or if you have a better idea, cool. Post away...
echaozh 16y ago
Perhaps not the source code, but definitely the detailed design with a lot of diagrams. And a working prototype.

by Hoff 16y ago

mseebach 16y ago

So why would source code for the "One Click" patent make any difference? It would take be all of ten minutes to make a perfectly functional implementation of "one click shopping" in an open source PHP cart which I could submit with the patent (or a day to build a simple cart from scratch if the open source-part is a problem). How would that make the patent less viable? Or are you arguing that only specific implementations could be patented? That, if we allow the premise of software patents in the first place, doesn't make any sense. Then we'd need measures to determine how much to change before it's not the same. I mean, to keep with the example of hardware patents, I'm sure a patent for a new crankshaft would also cover if it's 5% larger and made of a slightly different alloy. Re-implementing in an other language? What if it compiles to the same VM code? Then I'm getting the the business of generating and selling unique JVM languages. A new language for implementing the core of a competing algorithm? $1500, thank you.

Vivtek 16y ago
A patent for a new crankshaft might cover if it's 5% larger; it depends on how specific the claims are. There's a balance between too specific (thus useless) and too broad (thus not granted unless it's for freaking software); for machines this seems to be understandable, while for software the PTO just throws caution out the window, takes the money, and runs. The thing is, only a court decision will determine whether a patent covers a given use or not; I see a lot of documents written from both perspectives, and the real winners in the patent industry are lawyers (and, when all goes well and the case is sufficiently international, we translators get some nice money out of it, too).
Hoff 16y ago
Why? Because it (also) places indirect costs and additional onus onto the submitter of the patent; costs beyond the existing filing fees and legal fees. Because it also provides (some) tangible results for the community. (Now whether the covered source code is judged "good" or "bad" is another matter; there's certainly going to be some cases and some judgement required here around what's covered code, and whether or not the code is obfuscated, etc.) Because it means all of the patent-covered source code eventually gets exposed and licensed, which then gives an advantage to the community. Because even if your organization doesn't produce a product (and for whatever reason), you still have to produce functional source code. And because it means that most closed-source organizations probably won't choose to patent their software.
- mseebach 16y ago
  In your post you referred prominently to the one-click patent. I fail to see how your proposal would have even slightly curtailed that patent. > Because it (also) places indirect costs and additional onus onto the submitter of the patent; A developer for a day, as I demonstrated. > Because it also provides (some) tangible results for the community. > Because it means all of the patent-covered source code eventually gets exposed and licensed An implementation of an obvious process isn't a tangible, or even particularly valuable, result, as any member of the community could produce it in very little time. > Because even if your organization doesn't produce a product (and for whatever reason), you still have to produce functional source code. No, you just have to throw a $500 project on e-lance. Or about what you pay the lawyer who's writing the patent for you every hour. > And because it means that most closed-source organizations probably won't choose to patent their software. No, at best it would keep the stuff that's actually patent-worthy from being patented (and thus "out there", inspiring others, which is the whole idea behind patents), while the obvious crap gets patented (what does Amazon have to lose from spending a day writing a simple PHP cart?)..
  - Hoff 16y ago
    This scheme makes patents part of adding to open-source, and also uses the comparative "allergy" that many commercial entities have to open source. The folks seeking patents will have to decide if they want the patent enough to open-source their covered source code (later). Various firms certainly won't want that. Those that do go for the patents will then serve to produce open-source code. But if you don't think it'll work or if you have a better idea, cool. Post away...

echaozh 16y ago

Perhaps not the source code, but definitely the detailed design with a lot of diagrams. And a working prototype.