|Either Mathematica didn't really solve a TSP, or solving an NP-hard problem is becoming easier than I thought
It appears that Jordi is suggesting that Mathematica is itself an algorithm that proves N=NP. If he isn't, then his claim that Mathematica should be open to scrutiny (in the way that proofs must be) is unsupported.
Jordi's statistical exercise (interpolating surfaces to fit "scattered data"), with or without Mathematica, has no relevance to mathematical proofs that I can see (though I admit that I am not a statistician).
Did Wolfram Research insult the principle that scientific discovery (like better interpolation methods) should not be kept secret? Jordi doesn't say so.
The FSF's principles have no implications for private scientific research that I can see -- except in computer science (e.g. P vs. NP).
And what about the principle of a mathematical proof having rigour? Jordi denounces the inscrutable tool used to create the proof but not the negligent mathematician who accepts the partial proof as valid!
And why would anyone pay for a software package and then "tolerate" it?