suggestion for feature to improve document quality

[Tor Lillqvist]

> building lists of every word in the
> document would need implementing differently if it was to scale to large
> documents.

Or to languages more complex than English in their word structure.
Like languages that use inflections (for gender, number, function,
tense and whatnot). Basically, to be useful for such languages, this
new feature would need to understand the grammar (morphology?) of the
languages and be able to deduce the base word from an inflected word.
Not trivial at all. And I am pretty sure this is just not a finesse
that people could live without, but it would be the first bug somebody
writing in such a language and looking at the word statistics would
notice.

(The same holds for irregular verbs and nouns in English and other
languages, of course.)

--tml