uh, thought something looked strange, and if you're checking out the documentation at the webhelp link provided in the last post, there is a bit of an error in the explanation for:
!fieldname![1:4]
The result depicts the correct result 'sliced' from a string 'abcde' as being 'bcd' -- think of the character position indexes 0-5, and in the [1:4] expression, 4 is the 'upper bound' limit not included in the result. In other words 4 - 1 = 3 characters to return starting at position 1, or 'b'. (so the web doc explanation should not have included the fifth character)
You may have to play with it some...
Enjoy,
Wayne