I wrote a book-length research piece on collaboration and I still can’t spell collaboration correct on a consistent basis (I misspelled it in this sentence the first time through). Part of the reason is that I always use a spell-checker. The other part is because my spell checker (GNU Aspell) is really good. No matter how much I mangle a word, Aspell almost always manages to suggest the correct replacement and it’s usually the first option. The end result is that it’s more effort to learn to spell the word correctly than it is to correct it each time.
If my spell checker was less good and I was forced to read through the entire list options or, god forbid, type in the correct spelling by hand, I would know how to spell more words. I think that the lack of improvement in a users’ spelling ability over time may be one useful metric in evaluating the effectiveness and efficiency of a spell-checking software.
I think my complete stagnation in the swamp of bad-spelling is a testament to Aspell’s greatness.
While my English is so bad, that ASpell would coredump, I prefer not to use spell-checkers while writing in Polish. I try very hard to get on level where I’ll make no mistakes in short texts and few in an article. It’s just another thing to achive, like Python skills. ;-)
Ah, another note. In Polish we have few letters that are written in different way, yet, you pronounce’em in same way (no difference at all!) so, while góra (mountain) is ok, gura is not. Morze means sea, but it pronounce same as może (rz == ż) but then it means ,,maybe”. We have double chars for: ż == rz, ch == h, u == ó. Sometimes there’s no rule how to use them, so you’ll have to memorize it. Bah. ;p