Priti24:

If it's just people's names, then you don't need to do anything particularly heroic. You could have a small hash table of a reasonable range of roman numerals and look for a match in the correct location. Or, if the names are all formatted as in your examples, you could look for a comma followed by a regex. A simplified version would be would be something like: s/, I?V?I*//;. Extending to a larger range is left as an exercise for the reader.

Note: The regex will match some strings that aren't standard Roman numerals, and there's at least one other string it will match that it shouldn't. Generate *plenty* of test cases (especially degenerate cases) to tune your code against.

Have fun with it!

...roboticus

When your only tool is a hammer, all problems look like your thumb.