The term "
error
" should be generally avoided and substituted by the
term "
uncertainty
" of a result (except in such cases where indeed a real
error is involved in the measurement)
GUM recommends:
20