| |||||||||||||
FSF says: USPTO should publish guidelines excluding software patentsFSF says: USPTO should publish guidelines excluding software patentsPosted Sep 28, 2010 14:48 UTC (Tue) by sabik (guest, #70306)In reply to: FSF says: USPTO should publish guidelines excluding software patents by danielpf Parent article: FSF says: USPTO should publish guidelines excluding software patents
Yeah, I've read what you said. I'm still curious to know what you believe software consists of, if not algorithms. Nothing of what you've said actually answers that question.
The algorithm in a hello world program is as follows:
This is a fairly simple algorithm, having only one step and not taking any input, but it's an algorithm. It's a finite sequence of steps (just one, here) for solving a problem (such as it is). We can talk about its time complexity (which is constant) and space complexity (also constant). We can analyse whether it terminates (it does).
I don't know what you meant by "explicitly coded", but hopefully the above pseudocode satisfies that requirement: I've explicitly enumerated the one step in the algorithm. It is at the level of hierarchy customarily used by programmers, at least for simple tasks.
(Log in to post comments)
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 15:23 UTC (Tue) by danielpf (subscriber, #4723) [Link]
This simple program shows that not all of it is math at the level of programmer since "Hello Word!" is data. To claim the whole program is math you have to decompose the "Hello Word!" part at a sublevel not explicitly coded by the programmer where the characters are encoded for example in ASCII bits, which are then handled properly by a series of way more complicated system functions to be printed on screen.
I thought for LWN audience it would be obvious.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 16:13 UTC (Tue) by sabik (guest, #70306) [Link]
I'm afraid that you are simply mistaken.
Firstly, the algorithm that I presented is, in fact, an algorithm. No decomposition to bits or any other operation is required for it to be an algorithm - it's an algorithm the way it is.
Secondly, to quote Donald Knuth's letter linked above, "every algorithm is as mathematical as anything could be". It's not mathematical because it operates on numbers (although it might); it's mathematical because an algorithm is a mathematical concept. It's just as mathematical as a formula or a theorem or a construction in geometry or a fractal, none of which necessarily involve numbers.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 16:23 UTC (Tue) by danielpf (subscriber, #4723) [Link]
Did I say what you said is not an algorithm? No. I just said that not all the example program is math at the level the programmer would write it.
Did I say that an algorithm is not math? No.
Please read better what I say before replying otherwise the discussion becomes useless.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 16:29 UTC (Tue) by mpr22 (subscriber, #60784) [Link] For the sake of clarity, it might be best if you point out a non-trivial real-world program (ideally with publicly visible source code) and explain why you think it isn't mathematics at the level at which the programmer writes it. At this point, it seems to me that your position (not all programs are mathematics at the level of the programmer) depends on assumptions about the nature of programs and mathematics not necessarily shared by your audience.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 17:26 UTC (Tue) by danielpf (subscriber, #4723) [Link]
Your proposition is not a good idea: to single out a program for demonstrating a general trend is not a good way to convince sceptics.
I would prefer to use the known fact that the size of typical programs has grown following Moore's law while the functionality by far not. It should be clear that for a given amount of human time the produced software size (not counting data space) has grown without proportionality with the creative human work. Most of this growth results from using libraries and tools made by others.
If you say that calling a library is doing math, then I say that running the program itself, or even typing on the keyboard is also calling a library, and as we all use a computer we all do math all the time. This way of streching words may be fine in computer science circles, but make you lose any credibility by those in charge of making laws.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 18:10 UTC (Tue) by sabik (guest, #70306) [Link]
"Your proposition is not a good idea: to single out a program for demonstrating a general trend is not a good way to convince sceptics."
Technically that's not true, but fine. Answer my general question, then - what is software, if not algorithms?
If we accept that all software is algorithms and that all algorithms are maths, then it logically follows that all software is maths.
You seem to have already accepted the first part at least with respect to your own chosen example, the Hello World program. Indeed, you wrote: "Did I say what you said is not an algorithm? No."
You seem to have already accepted the second part universally. Indeed, you wrote: "Did I say that an algorithm is not math? No."
Therefore, we must logically accept the conclusion at least with respect to your own chosen example, the Hello World program, and concede that the Hello World program is maths.
The general case (about all software) depends on your answer to the question; what is software, if not algorithms?
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 18:35 UTC (Tue) by danielpf (subscriber, #4723) [Link]
Please read my original post *slowly*, digest, think, repeat, and don't come back with arguments again off the point, I will ignore them.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 19:20 UTC (Tue) by sabik (guest, #70306) [Link]
Well, you're already ignoring my arguments on point, aren't you, so at least I'll have the full set.
Your requests for re-reading your posts might be more productive if you pointed out where you think my understanding diverges from what you said. In what part have I misunderstood you?
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 20:02 UTC (Tue) by danielpf (subscriber, #4723) [Link]
Hint #1: I do not object that programs can be viewed as math when decomposed to its lowest levels. This is irrelevant to the discussion though.
Hint #2: I just say that using the slogan "software is math" is weak for arguing against software patents.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 20:46 UTC (Tue) by sabik (guest, #70306) [Link]
OK, sure. I don't see how that responds to my arguments, though.
1) The "Hello World" algorithm is an algorithm. No decomposition to lower levels or any other operation is required for it to be an algorithm - it's an algorithm the way it is.
2) All algorithms are math; not because they operate on numbers (although they might), but because an algorithm is a mathematical concept, like a formula or a theorem or a construction in geometry or a fractal.
Both of these points obey your hint #1; no decomposition to lower levels is involved in either of these two statements. In fact, you don't seem to object to either of them.
3) Combining 1 and 2, we can conclude that the "Hello World" algorithm is math. This is counter to your hint #2, but follows by logic that has been known to be valid for over two millennia. As a deduction, it is unlikely to be faulty (unless I'm mis-applying it in some way).
Now, I would really like to know what part of my argument is wrong. We seem to start from the same premises and reach opposite conclusions, so I have a feeling I'm missing something somewhere.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 23:29 UTC (Tue) by hppnq (guest, #14462) [Link] I think danielpf's argument against the "All software is mathematics" slogan is nicely illustrated by your insisting on its unequivocal truth. Read that Knuth letter again: the distinction between algorithms and software is a crucial ingredient of his argument against software patents.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 23:59 UTC (Tue) by danielpf (subscriber, #4723) [Link]
Let me try another way, perhaps you see light.
You perhaps just miss the first point I was pushing that arguing "software is math" is a weak argument in front of a patent legislator. The law makers had probably a different semantic interpretation of mathematics than yours when stating mathematical algorithms cannot be patented. Most probably they were thinking about the maths they learned at school, not the more recent maths elaborated during the 20th century, so "math" in their context covers classical theorems and algorithms, but not much more.
If you say any sequence of instructions is math, so cannot be patented (because finally computers are secondary on this matter) you will meet deep resistances for those wishing to patent business methods, drug synthesis, recipes, etc.
Saying any sequence of instructions is math leads you rapidly to name math about all the printed texts (each character is an instruction) and all music scores. Nature can be seen as a huge program because via its laws nature continuously executes "instructions" telling how each bit of the universe behaves. So everything is math, Pythagoras would appreciate, but no patent could be delivered anymore for material devices since they are also math.
This reduction to absurd (at least for common people) is often what happens when one takes the meaning of words in an strict sense ignoring the context.
Instead of arguing with arguments alien to a legislator, better invoke creative work for software which is sufficiently well protected by copyright law. My main final point is this, I hope it is clear.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 29, 2010 13:24 UTC (Wed) by nye (guest, #51576) [Link]
>Most probably they were thinking about the maths they learned at school, not the more recent maths elaborated during the 20th century, so "math" in their context covers classical theorems and algorithms, but not much more.
That interpretation isn't consistent with their decisions regarding the rest of the patent system. If the only reason to exclude mathematics from patentability is to avoid patenting *existing theorems*, then they may as well have excluded all inventions from patentability - after all, they didn't want existing inventions to be patented either.
>If you say any sequence of instructions is math, so cannot be patented (because finally computers are secondary on this matter) you will meet deep resistances for those wishing to patent business methods,
Good. Business methods should not be patentable.
>drug synthesis,
The list of instructions *already* isn't patentable. It may be copyrightable, and a device implementing those instructions may be patentable, but the instructions are not.
>recipes, etc.
Ditto.
>Saying any sequence of instructions is math leads you rapidly to name math about all the printed texts (each character is an instruction)
This is purely a straw man.
>and all music scores
Which are not, to my knowledge, patentable anywhere.
>Nature can be seen as a huge program because via its laws nature continuously executes "instructions" telling how each bit of the universe behaves.
In general, you seem to see some sort of reductionism where others do not - though you haven't yet coherently explained why you think any reduction is necessary - and then you extrapolate this perceived reduction ad absurdum for reasons which are not at all obvious to me.
Let's steer this into a slightly different direction, and take the example of MP3 patents. I'm picking the example because I can't think of anything which is more clearly a pure mathematical algorithm than data compression - do you disagree? And if so, can you explain why?
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 29, 2010 13:58 UTC (Wed) by danielpf (subscriber, #4723) [Link]
You seem skilled to misinterpret what I say and just derail the discussion.
You behave trollish so let us stop this discussion there. If you want to know, I am perfectly fine with Knuth's way to object software patents. Knuth doesn't need to invoke "software is math" to make his points. So let us stop boring the LWN readership with vacuous statements.
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 29, 2010 14:24 UTC (Wed) by nye (guest, #51576) [Link]
>You seem skilled to misinterpret what I say and just derail the discussion.
>You behave trollish so let us stop this discussion there.
Whaaaat??
That attack was unwarranted. I was trying my best to understand your position, despite your total inability to coherently express it.
Given that this isn't your first personal attack in this thread on people trying to understand you, I can only assume that your outlandish nonsensical claims are simply a (sadly successful) attempt to troll this article.
*plonk*
FSF says: USPTO should publish guidelines excluding software patents Posted Sep 28, 2010 16:40 UTC (Tue) by sabik (guest, #70306) [Link]
My apologies - I seem to have omitted the key syllogism from my argument.
OK, labelling my "Hello World" algorithm as H, we have:
1) H is an algorithm.
2) All algorithms are mathematical.
Therefore:
3) H is mathematical.
|
|||||||||||||