The words

The probe answers what is in here. This part is the other half: taking what you want out.

Four words so far — rows, first_present, take and where — and each one had to earn its place against a rule that is deliberately hard to satisfy:

A word belongs in fathom only if removing it makes one of the fixed questions unanswerable on at least one corpus file. No word is added without naming the question it answers and the file that proves it.

Every word in this part is shown in both languages, because the premise of the project is that they read the same. That is a claim, not a courtesy, and it is checked rather than asserted: the two spellings come from two independent implementations and a harness diffs what they return on real documents.