The REPP is a system for rewriting strings via chains of regular expressions, with some cute iteration and modularity features thrown in. It is used primarily to harmonize punctuation affixation and split sentences into tokens. Its output is a sequence of (rewritten) tokens. For downstream applications it is desireable to know the character positions in the original string corresponding to the rewritten tokens. Due to insertion, deletion, and modification of characters by the rewrite rules, this is not quite a trivial problem. For arbitrary rewrite systems, it can be very unclear what the "correct" character positions to associate to output tokens are, e.g. consider this improbable RPP.