help make it faster please

[bearophileHUGS at lycos.com]

Thank you Bengt Richter and Sybren Stuvel for your comments, my little
procedure can be improved a bit in many ways, it was just a first
quickly written version (but it can be enough for a basic usage).

Bengt Richter:
>good way to prepare for split

Maybe there is a better way, that is putting in just the accepted
letters and accented letters, instead of the not accepted symbols.


>I suspect it's not possible to get '' in the list from somestring.split()

You are probably right, the algorithm used is different if you don't
give a splitting string:
>>> ' abc  a '.split()
['abc', 'a']
>>> '.abc..a.'.split(".")
['', 'abc', '', 'a', '']


>does that beat the try and get versions? I.e., (untested)

Yes.


>countDict[w] = countDict.get(w, 0) + 1

I think the if-else version is the fastest, I have tested it a long
time ago... You can easely do a speed test to see if I am wrong.

Bear hugs,
bearophile