>>Very nice, but...There are at least two problems with this technique. First, it requires that all the values be integers. You cannot implement this with characters or floating point numbers. Second, it requires that they be unique. If they're not, there will be at least one instance where the sorted array doesn't contain a value. IMHO, any sorting technique of merit should be able to properly arrange the values regardless of type and range.
>>
>
>Yes, you're right George, the Quick got my vote as fastest on average, and I verified it in my books, as I noted elsewhere - I have quite a number of books on sorts, searches, and trees available here, using C/Pascal as the example lingos. BTW, the "pigeonhole" got a mention in one of them, but was shrugged off as not valid for general use.

Thanks, Bruce. The biggest problem with the Quick is the implementation. It is not easy to understand. Right now, I'm trying to work one up and it's a major PITA. Hopefully, I'll have it done during lunch and post it (with the others) here.