My Work Log

Well how would a person approach the task of estimating an application.  It simply is not as straightforward as picking up the application, downloading templates, applying the methodology and saying kaput ….that’s the estimate.  There are so many other factors that come to play.  For eg.:

• Who is doing the estimation ……..is he a techie or is he a BA.
• What is his mandate …….has he been told to be conservative …….or to go overkill!!
• His skill level on the methodology
• What rules he wants to bend
• How was his initiation in the methodology - was he formally trained …or was it trial and error.

My experience on doing FP estimation has been so varied ………..that to date I have not seen a real pure implementation of FP.  Every time I get down to do it …….my co-workers would bring in one form of corruption or the other. 

Now for the uninitiated ………FP or Function Points is a methodology of sizing either a new development or an existing application.  Once the sizing is done …….the result gives you the Function Point Count for the application.  This is translated using a ‘productivity factor’ into Man-hours of effort.  Talking more of the productivity factor ….it simply reflects the technology used to translate the application into the underlying software. 

But to arrive at the initial Function Point count ………you do not take technology into consideration.  FP estimation is purely done keeping the end user into account ………and it strictly defines a boundary around the application …………inside which ……no matter what happens ….we should not take it into account.  So its more like the user interacts with a black box.  Whatever functions come out of the black box to the user are only what are counted ………and not what intrinsic functions happen within.  This is one rule which usually takes a beating …….when the FP estimation is done by a technical person ……….who wants to emphasize the many technical functions which happen within the black box. 

The next aspect is when we talk of functions ……………each function should be seen independently.  Which means that if the black box does ten things to the user - then the no. 10 is arrived at based on the fact that each of those ten things are independent and can function without being dependent on the other 9.  This is another rule which gets bent so often.  People count part functions as distinct functions. 

Thus at the end of the day there is more confusion to the methodology than meets the eye.  FP is an estimation methodology which has a certifying authority called the IFPUG ……….which conducts the CFPS exams (Certified Function Point Specialist).  Clearing this exam is a challenge with the pass mark being 90%.