Both approaches have their advantages...you might consider doing a big-file scheme with some sort of index or allocation table at the start, so that files can be quickly and easily found by their offsets...If you could somehow interface this with the archive memory, it might be fun. :-) --Cliff Biffle