Fine, but since his tools are used to prove mathematical theorems, how can we formally be sure about the correctness of Mathematica statements if Mathematica is just black box (I agree though this is more a matter of principle). Additionally, given all the movement in Open Science and Open Access, Mathematica and the other jewels could be very well part of it. E.g. take all the software produced at CERN to analyse LHC data. It's all Open Source. (and the scientific outcomes are all Open Access) In principle
since his tools are used to prove mathematical theorems, how can we formally be sure about the correctness of Mathematica statements if Mathematica is just black box (I agree though this is more a matter of principle).
Yeah, since this closed-source software doesn't contain any bugs that give you the wrong answer.
Well, a blackbox proof is useful -- a doubter can submit something that generates an error.
If I were him, I'd strike a deal with Google.for all the formulae they've come across.
For that matter, there are other math engines, and probably an alternate development one in-house, and you can test them against each other with random inputs until you find a discrepancy.
I've done that myself with algorithms several times -- the "real" one against another, less-efficient and different one.
They debug each other. It's my experience code inspection is a useful but inferior tool to this approach (for non-malicious discrepancies.)
Why not open source wolfram alpha? (Score:3, Insightful)
Re: (Score:0)
Because he runs a business not a hippy commune. Go away little freetard.
Re: (Score:0)
Fine, but since his tools are used to prove mathematical theorems, how can we formally be sure about the correctness of Mathematica statements if Mathematica is just black box (I agree though this is more a matter of principle). Additionally, given all the movement in Open Science and Open Access, Mathematica and the other jewels could be very well part of it. E.g. take all the software produced at CERN to analyse LHC data. It's all Open Source. (and the scientific outcomes are all Open Access)
In principle
Re: (Score:0)
since his tools are used to prove mathematical theorems, how can we formally be sure about the correctness of Mathematica statements if Mathematica is just black box (I agree though this is more a matter of principle).
Yeah, since this closed-source software doesn't contain any bugs that give you the wrong answer.
Re:Why not open source wolfram alpha? (Score:1)
Well, a blackbox proof is useful -- a doubter can submit something that generates an error.
If I were him, I'd strike a deal with Google.for all the formulae they've come across.
For that matter, there are other math engines, and probably an alternate development one in-house, and you can test them against each other with random inputs until you find a discrepancy.
I've done that myself with algorithms several times -- the "real" one against another, less-efficient and different one.
They debug each other. It's my experience code inspection is a useful but inferior tool to this approach (for non-malicious discrepancies.)
Re: (Score:1)
God what an idiot I am. He's got his own world-class web crawler and data analyzer.