qweasd |
11-14-2011 11:39 AM |
Quote:
Originally Posted by tangle
(Post 4523713)
So, you go through every line of code in every piece of mathematical software you use?
|
I go through some every now and then. When looking for correct ways to generate pseudo-random rationals, for example, I personally read through the code in GNU binutils, awk, perl, and others I am forgetting now. Every time I have to do this, I am pleasantly surprised. Reading Paul Eggert's code was not only remarkably easy, but also made me feel enlightened :)
But of course, even if I couldn't read C to save my life, I'd still have the same attitude. I do trust the world-wide community of hackers and mathematicians to review the code. I don't trust a proprietary software vendor for a second. These are the same people who root your systems and spy on your files.
And even beyond the issues of personal trust, it's just hilarious that people (scientists!) are willing to throw away all of their notions about the integrity of the scientific research for a shiny piece of software. It's not peer-reviewable, and that should be the end of the argument. A good word for using this software for purposes other than entertainment is "corruption". I hope you can see how this becomes the ultimate joke when applied to math: the only thing we have to show for our efforts are proofs. A valid proof is what turns a statement into a mathematical result. In this sense, MATLAB doesn't produce any math at all.
|