crazy semantics in The Wolfram Language

[Dan Drake]

See the screenshot at
https://plus.google.com/+DanPiponi/posts/fF9xHcxo9Ze .

Quite a violation of the principle of least surprise...


Dan

--
--- Dan Drake
----- www.math.wisc.edu/~ddrake/
-------

[maldun.fin...@gmail.com]

Actually from what is shown, it's not a problem with the principle of least suprise, it's simply wrong:

the sequence f is defined by f[0] = 1, f[n] = n for n > 0, thus summation over n =0,…,m gives

Σ f[n] =  ½ m(m+1) + 1

It appears that the symbolic summation ignores the special definition of f[0] = 1,

and only cares about the definition f[n_] = n, and thus it gives a wron result, in the symbolic case,

and gives the right one for the numeric value, where (at I suppose) f[n] is evaluated explicitely.

Nevertheless Mathematica does it wrong.

[maldun.fin...@gmail.com]

Sorry for replying agein, but I saw that in one of the comments from your link it seems

that Maxima makes the same mistake.

- Stefan

On Friday, December 19, 2014 4:17:41 PM UTC+1, Dan Drake wrote: