Formal Languages and Automata Theory papers at this level usually run for 2 to 3 hours and carry 60–70 marks. A common layout is one compulsory question (often 20–25 marks, built from short parts across the whole syllabus) plus any three of five or any two of four optional questions; some schools instead set five or six equal questions and ask for four. Each question is normally split into parts worth 3 to 10 marks. Work out your minutes per mark at the start: on a 70-mark, 3-hour paper you have about 2.5 minutes a mark, so a 10-mark conversion deserves about 25 minutes.

The questions have a clear personality. About a third are definitions and comparisons: define a DFA or PDA as a tuple, distinguish DFA from NFA or Moore from Mealy, state the pumping lemma, describe the Chomsky hierarchy, explain the halting problem. The rest are constructions and traces: design a DFA for a described language, convert an NFA or ε-NFA to a DFA, minimise a DFA, convert between finite automata and regular expressions, derive a string and draw its parse tree, show a grammar is ambiguous, convert a grammar to Chomsky Normal Form, design a PDA or Turing machine and trace it on an input. Finite automata and context-free grammars usually carry the most marks; Turing machines and decidability more often appear as shorter parts.

A calm strategy: answer the questions with the most tables and traces first, because they reward method and have one right answer. In every construction, show the working (each subset row, each partition, each derivation step, each instantaneous description), label the start and accepting states, and finish with a check on one accepted and one rejected string. For proofs with the pumping lemma, write the full game: assume regular (or context-free), let $p$ be the pumping length, choose $w$, consider every split, pump, and conclude with a contradiction. Leave a few minutes to recheck the cells of any table.