> Then they should make it a word limit and give you stop words for free (a, the, and, etc).
People would work around that by dropping punctuation so the algorithm would not see word boundaries. Or by writing in German...
People would work around that by dropping punctuation so the algorithm would not see word boundaries. Or by writing in German...
A better (but more opaque) system might try to estimate the "quantity" of semantic content in a language-specific way.