When SSA-NAME3 matches a person name in Greek that starts with

P

with its equivalent name in Latin-1, it returns a lower score than expected. This is because the

n3csgr2.dat

option incorrectly converts all instances of

P

in the name of a person to

R

.

The

FMTOVERRIDES

function keyword doesn't work on a testbed or population file.

When you perform a search without configuring a value for the Group control parameter, you get an incorrect score.

When the options

CONCAT

and

RAW

perform a raw comparison of concatenated words between two person names, it might return a lower score than expected.

The

NM3KEYSIZE=12

parameter that specifies the key length for fields is not specified in the code key definitions in standard populations.

If you use the options

ORIGWORD

along with

FMTINIT,1

, the

ORIGWSCR

and

ORIGWTHR

options result in a higher score than expected when matching two names with initials.

The

ssan3_get_keys

function doesn't work as expected. When the values for the

CODE

field have multiple leading zeroes, the function generates identical keys instead of different keys.