Syntax Directed Translation Questions and Answers

Answer: C

Correct answer: the type of an expression is synthesised from those of its operands

The type of a sum follows from the types of its two operands, which follow from theirs, all the way down to the identifiers and constants. That is a purely synthesised computation, so type checking fits an S-attributed definition and can be done during a bottom up parse.

Answer: D

The rule detects the mismatch and inserts an explicit conversion so that later phases see consistent types. Whether a coercion is permitted at all is a decision of the language definition. Some languages forbid every implicit conversion and require an explicit cast.

Answer: A

The place attribute names where the value lives, usually a compiler generated temporary. The code itself is either another attribute or is emitted directly as a side effect. Which choice is made affects whether fragments can be reordered afterwards.

Answer: B

A counter guarantees a distinct name for each intermediate result. Reducing the number of temporaries is left to later optimisation rather than to the translation itself. Keeping the translation simple makes it far easier to verify as correct.

Answer: C

The multiplication produces one temporary and the addition a second, giving two in all. Each binary operator in the tree contributes exactly one temporary under this scheme. Counting operators is therefore enough to predict the number of temporaries.

Answer: D

Each operator becomes one instruction, so five are emitted. The number of temporaries introduced is also five, one holding the result of each operation. Later optimisation may eliminate several of them by reusing registers.

Answer: A

Passing the destinations down the tree lets each subexpression emit a jump directly to the right place. This avoids computing a value that would immediately be tested and discarded. The saving is substantial in code full of compound conditions.

Answer: B

Correct answer: the destination of a jump is not known when the jump is emitted

Translating a conditional or a loop in one pass means emitting the jump before the code it jumps over has been generated, so the destination cannot yet be known. The jump is emitted with the field blank, its position is recorded on a list, and the list is patched when the label is finally reached.

Answer: C

Merging two such lists is how the true exits of a subexpression are combined. When the destination becomes known every instruction on the list is patched in one sweep. The lists are merged as subexpressions combine, which keeps the cost linear.

Answer: D

Values built from parts flow upward while context supplied from outside flows downward. Recognising which applies is what determines whether the definition can be S-attributed. A single downward flow anywhere forces the L-attributed treatment.