I want to compare the current code for naming classical groups, i.e. the function RecogniseClassical, to the corresponding magma code. If there are discrepancies with the magma code, then I'd simply use the magma version.
RecogniseClassical is designed to be used with natural matrix groups. Until now we pass representatives of generators of a projective matrix group. Is it possible, that we pass matrices, such that taken projectively they generate a simple group, but as a matrix group they do not generate the full corresponding quasi simple group, that is parts of the center could be missing? If that can happen, then I suggest the following workaround: after determining the forms that we leave invariant, create a new group to whose generating set we add generators of the center of the quasi-simple group corresponding to the form. That one is then guaranteed to contain the quasi-simple group.
Also, the function should probably be renamed to NameClassical.
I want to compare the current code for naming classical groups, i.e. the function
RecogniseClassical, to the corresponding magma code. If there are discrepancies with the magma code, then I'd simply use the magma version.RecogniseClassicalis designed to be used with natural matrix groups. Until now we pass representatives of generators of a projective matrix group. Is it possible, that we pass matrices, such that taken projectively they generate a simple group, but as a matrix group they do not generate the full corresponding quasi simple group, that is parts of the center could be missing? If that can happen, then I suggest the following workaround: after determining the forms that we leave invariant, create a new group to whose generating set we add generators of the center of the quasi-simple group corresponding to the form. That one is then guaranteed to contain the quasi-simple group.Also, the function should probably be renamed to
NameClassical.