A FORTRAN program has been written which inputs amphibole analyses, and outputs complete IMA-approved names, along with tabulated analyses, formula units, and IMA classification parameters. Several important corrections, additions, and amendments to the IMA scheme were found to be necessary. Two newly approved names have been incorporated—sadanagaite (ferro-pargasitic amphibole with Si < 5.5), and nyboite (alkali amphibole with (Na+K)_A ⩾ 0.5, Fe³ ⩽ Al^VI, Si < 7.5). The name anophorite (alkali amphibole with (Na+K)_A ⩾ 0.5, Fe³ > Al^VI, Si < 7.5) has been reinstated to complement nyboite, with the approval of the Amphibole Subcommittee of the IMA. Results of computer processing of several hundred analyses, which together cover the range of IMA-approved names, are summarized. These indicate substantial differences between IMA and previous nomenclature. The IMA-favoured procedure for reallocating total Fe in microprobe data is quite effective, except in kaersutites. Over half the analyses processed have anomalously high cations (e.g. Ca > 2), or cation totals (e.g. T + C + B > 15), commonly reflecting low H₂O determinations. These prevent strict obeyance of the IMA site occupancy calculation rules. This problem can often be overcome by recalculation free of H₂O, but other cases require modifications to the IMA rules. In particular, ‘misclassification’ of certain sodic-calcic amphiboles as edenite follows from transferring large excesses of C cations to the B site.