Writing to unique files

Hi, Does anybody have any neat tricks for a good way to choose unique sequential filenames in the following format?: abc-000001.xyz abc-000002.xyz etc .... in other words, what fast methods are there for creating a new file which will be the next in the sequence in a directory where the files follow the above format? Is there any quicker way than laboriously reading in every directory entry, extracting the number and then determining the largest value and adding one!? Also, how can one guard against potential conflicts (ie two processes running at such a close time that they both select the same filename)? Many thanks in advance! Louis