On Saturday 17 October 2009 20:10:44 Jorge Llambías wrote: > Maybe the database should contain N matching pairs, and for a given > test randomly select n < N pairs, randomly ordered. That way we get > lots of different versions of the question relatively cheap. Sounds good. I'll make another table for the matching pairs. Pierre -- lo ponse be lo mruli po'o cu ga'ezga roda lo ka dinko