In the usually simple tasks I give to REBOL, I've found that more often than not it helps to start writing the code backwards - that is, write the "output" function first, the one that will actually write the HTML pages. Forget about the mechanics of reading in the directories and filenames, and think instead of the most convenient form (ie data structures) you would like the information when you create the HTML pages. Make this a parameter(s) to the output function. THEN your main function will be about reading in the data and populating this desired data structure. In other words, I've found it helpful to break a simple task (such as yours) into two stages: reading the data into a REBOL structure, then writing the data out from this structure. I don't even try to combine these two steps anymore. It's faster just to separate these two things from the beginning. - Michael Jelinek -- To unsubscribe from this list, please send an email to [EMAIL PROTECTED] with "unsubscribe" in the subject, without the quotes.