Panofoot:

What is partitioning?  If one goes to a desktop or on-line dictionary and looks up the word "analysis", chances are that the first listed definition will state basically that "Analysis is the partitioning (i.e., dividing up) of an entity into its parts, and then examining the parts and their interrelationships".   If I were to rigorously analyze a physical entity like, for example, a coffee maker, I would take it appart; look at all the individual parts; and - most importantly, evaluate  how each part interrelated with other parts.

(Note:  Once we properly partition a system, we can then decompose each function (i.e., part) downward.) 

Tony

Pano:

Great questions!   In function/process  modeling, the "parts" are typically called functions on data flow diagrams, processes on buisness process maps, and use cases on a use case diagrams.  (Key point:  There is no logical reason to refer to the same thing by different names; don't be mystified by the mechanism (i.e., whether or not the process/function is automated or not))   

If you are a data flow diagrammer, there definitely is a formalized approach to partitioning a system.   Key elements to that approach are:

*  A logical, natural partitioning of the system.  (For further details on what a logical, natural partitioning is and how to do one, please refer to my previous posts on such.) 

*  Further refinement of the natural, logical partitioning based upon evaluation of each discovered function's interface complexity (i.e., if a function has too many inputs and outputs, it is a clear sign that further partitioning of it is needed).

Use cases do not employ a formal approach to partitioning.  The default approach to formal partitioning is called brute-force partitioning, or forced, artificial partitioning (which I have also discussed in previous postings).   Business process mapping techniques espouse partitioning largely based upon sequencing and flow of control considerations.   Such an approach is mainly appropriate for more smallish efforts. ( I have discussed why in previous postings.) 

Pano, as I  stated, you ask some great questions!   Reason:  As analysis is literally defined as partitioning, unless we primarily position analysis techniques relative to each other in terms of their approach to partitioning (or lack thereof), we will get mixed-up notions as what technique to use when.  

Tony

There is something in your comments for everyone.

When you get good at someting and continue to get good results from using it, it's easy to stick with what you know.

However it's important to continue to expand your knowledge and skills.  Sharpen the saw.

I've spent most of this year doing just that and feel I have elevated myself to a new leval of capability.  I recommend it to everyone who feels that they are good at what they do.