Unit 2 - Numbers, Arithmetic, Expression trees

Question 1

A notable exception is Lisp where multiplication can
take any number

Answer 1

of operands (ie it has a mixed arity)

Question 2

Java automatically inserts integers to float conversions where necessary but *****
does not.

Thus in **, we have to write ** where in Java 2 is sufficient.

Answer 2

Ada

2.0

Question 3

Haskell does not insert any conversion, but instead it

Answer 3

works out from the context that 2 is a float rather than an integer.

Question 4

Integer CPU types usually come in two flavours: (and ranges)

Answer 4

singed and unsigned.

An unsigned 8-bit integer, has the range 0 . . . 255.

A signed 8-bit integer type has the range −128 . . . 127.

Question 5

A floating-point overflow happens only when

Answer 5

the exponent does not fit in the available space.

Question 6

An overflow upwards (ie for a very large positive number) is

Answer 6

rounded to positive infinity (∞)

Question 7

an overflow downwards (ie for a very large negative number) is

Answer 7

rounded to negative infinity (−∞).

Question 7

Positive and negative infinities are

Answer 7

are special values of floating-point types.

Question 8

A floating-point underflow is a situation where the result of an operation is

Answer 8

so close to 0 that the exponent is a very large negative number.

Question 9

An underflow usually results in

Answer 9

rounding the number to 0.

This does not happen easily because floating-point types are most accurate near 0.