[SciPy-dev] Stats t-test broken on 1D input

[Ed Schofield]

> Re-writing something in a completely different language isn't
> transformative enough (especially given its short length) to make the
> resulting code no longer copyrighted by Numerical Recipes?
>
> I am of course NAL, but I find that surprising.

Well, this is my interpretation of the book's License Information  
section (at http://www.library.cornell.edu/nr/bookcpdf/c0-1.pdf).   
Translations of novels into other languages are also considered  
"derivative works" and require permission of the copyright holder.   
But it's hard to draw the line between idea and expression in  
computer code.  If it's really a straightforward implementation of a  
well-known algorithm, like Robert argues, then perhaps we'd just want  
to remove the references to NR, rename the variables, add some nice  
comments -- and stamp out the bugs ...

-- Ed