People's names are a key in our database structure, thus different people must have different names. If they have the same name, we have to use something to make them different, in these cases we use Roman numerals.
We make two different types of uses of Roman numerals:
1) When they are not in parentheses, like Lloyd Ahern II, this is an actual part of someone's name. In Lloyd's case, his father has the same name, so the familial relation is represented by the II. When some people more commonly or very consistently use "Jr." we do not use a roman numeral as the junior notation makes the names different.
2) When they are in parentheses, that is an internal notation for the database. We may have a costume designer, a director, and an actor all named John Smith, none of whom are related. If we did not separate them by use of parenthesized Roman numerals--the (I) and (II), etc., that you see--their credits would all be combined, meaning there would be one John Smith with all of their credits instead of three John Smiths with the proper credits for each.
There are no hard and fast rules as to how they get ordered, though we try to use one or more of the following standards: who is more famous, who came first chronologically, whose job is more popularly looked up (i.e. when people are looking up a name, they're more likely to be looking for information on a director or actor than a production assistant or key grip with the same name), who was added to the database first.
To avoid confusion among our users and to prevent potential loss of data due to unnecessary corrections, we do not change or reorder roman numerals once they have been assigned, except as dictated by internal technical requirements.