In Figures A and B I've set up two simple procedures for random data access. Given a strKey both return the country name associated with the key. The Country table has 252 countries in it. I've stored it as both a dictionary and an array—the dictionary code is four times as fast.       In Figures C and D I've created a similar scenario. This time, though, I loop through the data set, loading a HTML select (AKA combobox) with the list of companies. There's no reason to use a dictionary here. You could if you didn't want to mix arrays and dictionaries in the same page or project, but there's nothing wrong with mixing them, provided you wrap your array functions well.        Figure E details the results of a more elaborate performance test. This test, downloadable from the MIND site, uses Manohar Kamaths' ASPTime.dll (see http://www.kamath.com) to act as the millisecond (± 5) level stopwatch. The results speak for themselves. (Anyone hear falcons?)       The iteration numbers are low because I wanted to be consistent across all tests. Sure, loading up 100 Kb of records from a simple table is quick, but just look at the time for a single array lookup in my deepwide array—24 seconds! For just what these tests are composed of—come find out, the code is free to all.