3 ms·

What if a company programs directly in assembly? Should that make them ineligible for patents? And if assembly is only allowed when it's the original language,

by eurleif 14y ago

What if a company programs directly in assembly? Should that make them ineligible for patents? And if assembly is only allowed when it's the original language, how would you go about proving that? What about code that's been compiled, then decompiled back to the original language? Or just put through an obfuscator? Basically, I think the nature of code makes it impossible to set up useful standards.

lutusp 14y ago
> What if a company programs directly in assembly? Should that make them ineligible for patents? Not at all. If source code is required, they would simply provide it. I think the idea is to be as transparent as possible, so any assemblable or compilable source should qualify. I should add that I wrote a lot of assembly code years ago -- Apple Writer (http://en.wikipedia.org/wiki/Apple_Writer http://en.wikipedia.org/wiki/Apple_Writer) was entirely assembly code -- so I don't think of it as unreadable or even vague. But this might not be very important -- the intent of requiring "working code" may only be to prevent software patents that are too broad (and that's discussed in the article). So this talk about source versus executable may miss the point.
- eurleif 14y ago
  Sounds like this is a moot point, but I'm still curious: How would you legally distinguish between programmer-written assembly, and assembly disassembled from compiled machine code? Without doing that, it's pointless to require source code, since disassembler-generated assembly is no more useful than the machine code it was generated from. And it seems like no matter what rules you set for how the assembly should look, someone could change it just enough to comply with those rules, while still not being very useful.
  - lutusp 14y ago
    > How would you legally distinguish between programmer-written assembly, and assembly disassembled from compiled machine code? The same question could be asked about certain languages that possess this "de-compilation" property, including Java. And the answer is, if someone understands the code well enough, (s)he can sprinkle comments and plausible-sounding function and variable names and pass the result of as the original. > Without doing that, it's pointless to require source code, since disassembler-generated assembly is no more useful than the machine code it was generated from. Yes, to a programmer, but to a lawyer it makes a world of difference. An invented source code file would have two traits of use as evidence -- it would be byte-for-byte identical with the real source, but it would have gaps in the reasoning expressed in the comments, and it might show implausible function and variable names. Also, apart from providing a basis for a fraud allegation, requiring a source file takes more effort than simply providing an executable. > And it seems like no matter what rules you set for how the assembly should look, someone could change it just enough to comply with those rules, while still not being very useful. When all is said and done, I agree. Although such a requirement places obstacles in the path of cheating, it's not any real barrier to someone who intends to produce what at first glance seems to be the original source code. Also, we don't know if this requirement even exists, which was the reason for my original question. Requiring source might not even be in the minds of those making this proposal.

by eurleif 14y ago

lutusp 14y ago

> What if a company programs directly in assembly? Should that make them ineligible for patents? Not at all. If source code is required, they would simply provide it. I think the idea is to be as transparent as possible, so any assemblable or compilable source should qualify. I should add that I wrote a lot of assembly code years ago -- Apple Writer (http://en.wikipedia.org/wiki/Apple_Writer http://en.wikipedia.org/wiki/Apple_Writer) was entirely assembly code -- so I don't think of it as unreadable or even vague. But this might not be very important -- the intent of requiring "working code" may only be to prevent software patents that are too broad (and that's discussed in the article). So this talk about source versus executable may miss the point.

eurleif 14y ago
Sounds like this is a moot point, but I'm still curious: How would you legally distinguish between programmer-written assembly, and assembly disassembled from compiled machine code? Without doing that, it's pointless to require source code, since disassembler-generated assembly is no more useful than the machine code it was generated from. And it seems like no matter what rules you set for how the assembly should look, someone could change it just enough to comply with those rules, while still not being very useful.
- lutusp 14y ago
  > How would you legally distinguish between programmer-written assembly, and assembly disassembled from compiled machine code? The same question could be asked about certain languages that possess this "de-compilation" property, including Java. And the answer is, if someone understands the code well enough, (s)he can sprinkle comments and plausible-sounding function and variable names and pass the result of as the original. > Without doing that, it's pointless to require source code, since disassembler-generated assembly is no more useful than the machine code it was generated from. Yes, to a programmer, but to a lawyer it makes a world of difference. An invented source code file would have two traits of use as evidence -- it would be byte-for-byte identical with the real source, but it would have gaps in the reasoning expressed in the comments, and it might show implausible function and variable names. Also, apart from providing a basis for a fraud allegation, requiring a source file takes more effort than simply providing an executable. > And it seems like no matter what rules you set for how the assembly should look, someone could change it just enough to comply with those rules, while still not being very useful. When all is said and done, I agree. Although such a requirement places obstacles in the path of cheating, it's not any real barrier to someone who intends to produce what at first glance seems to be the original source code. Also, we don't know if this requirement even exists, which was the reason for my original question. Requiring source might not even be in the minds of those making this proposal.