On a Divide-and-Correct Method For Variable Precision Division
Described in this paper is a divide-and-correct
method for variable precision division in digital 
computers.  Unlike the earlier methods of Stein and
Pope, the present method uses a suitably rounded 
form of the normalized divisor for getting an estimate
of the quotient characters.  This results in a 
correction of at most plus or minus one to the estimate,
to obtain the exact quotient character.  It 
is believed that this method will be widely applicable
for division operations in variable word-length 
character-oriented machines.
CACM March, 1965
Krishnamurthy, E. V.
