Syntax Directed Translation Questions and Answers

Answer: A

Correct answer: a set of semantic rules relating the attributes of its symbols

A syntax directed definition pairs each production with rules saying how the attributes of the symbols in it are related - the value of the left side from those of the right, or the reverse. It says what is to be computed without saying when, which is what distinguishes it from a translation scheme.

Answer: B

Correct answer: a value associated with an occurrence of that symbol in a tree

An attribute belongs to a node rather than to the symbol in the abstract, so the same nonterminal appearing twice in a tree carries two independent sets of values. Attributes hold whatever the translation needs - a type, a value, a piece of generated code, an address.

Answer: C

Correct answer: the attributes of that node's children and of the node itself

A synthesised attribute flows upward, computed from below. That direction is what makes S-attributed definitions so convenient for bottom up parsing: by the time a production is reduced, every child has already been reduced and its attributes are available on the stack.

Answer: D

Information flows downward or sideways across the tree. Inherited attributes express context, such as the type declared at the head of a declaration being passed to each name in the list. Without inherited attributes that construct cannot be handled in one pass.

Answer: A

Correct answer: A-2, B-4, C-1, D-3

The direction follows from where the information comes from. An expression’s value and a code fragment’s width are both built up from the parts below, so they are synthesised. A declared type must be passed down to the list of names it applies to, and a nesting level comes from the enclosing block, so both are inherited.

Answer: B

Correct answer: every attribute in it is synthesised

S-attributed definitions use no inherited attributes at all, so information flows only upward. That makes them exactly what a bottom up parser can evaluate during the parse, with the attribute values held on a stack parallel to the parsing stack and no extra machinery at all.

Answer: C

Every attribute is synthesised, so the children are complete when the parent is reduced. Yacc-style value stacks implement exactly this evaluation with no extra machinery. Nothing beyond the parallel stack is needed to realise the whole definition.

Answer: D

Correct answer: the left side's inherited attributes and symbols to its left

The restriction is to information already available when the symbol is reached in a left-to-right traversal - what came down from the parent, and what has been computed from the siblings already passed. That is precisely what makes a single depth-first left-to-right pass sufficient.

Answer: A

Correct answer: 1 and 3 only

Statements 1 and 3 are correct: an S-attributed definition has no inherited attributes at all, so it satisfies the L-attributed condition vacuously, and L-attributed definitions evaluate in one depth-first pass. Statement 2 reverses the containment - an L-attributed definition may use inherited attributes, which an S-attributed one by definition does not.

Answer: B

Both are true and the reason gives the mechanism, and the containment is strict. An L-attributed definition passing a declared type down a list of identifiers uses an inherited attribute and so lies outside the S-attributed class.