At work the other day I used the following descriptions for describing the 4 types of clustering that we use in a SAS environment. I think that they are really useful short descriptions of each type so noting here for future use.
An interesting problem at work a while ago came about when a colleague was trying to perform a calculation on two numbers but instead of getting back the result the program was returning a NULL value. It turns out that the reason for this is that the number was so big that his machine was not able to handle it. Not ideal. So how can we pre-empt this and ensure that the calculation that we want to do is achievable within our environment?