ragged arrays vs matrices in data read

it's hard to tell from the format of a data file alone whether an object is supposed to be a ragged array (which _might_ actually be balanced, i.e. "non-ragged", in any particular run) or a matrix. _If_ the data lines are different lengths, then it must be a ragged array, but if they're not then it might be either. One solution would be to use TPL info to resolve the difference ...

Start by locating the data-reading entry point and reviewing how TPL information is available to it. Determine the intended rule for distinguishing ragged arrays from matrices when all data lines have equal length, then document the decision and add coverage for both ambiguous and unequal-length inputs.

Written by the indexing model from the issue text.