Using Function Points for Maintenance Effort

Basic Principles

Removing function points is cheaper than adding them.
Adding n FPs to a N FP system is relatively cheap when n < < N.
The 14 value adjustment factors don't apply to maintenance changes.
The new value adjustment factors are applied, individually, to each Transaction Function Type or Data Function Type.

Maintenance Impact Ratio (MIR)
Subject measures of the impact of a change.
Used as a multiplier for the Unadjusted Function Point counts.

MIR is 5-valued scale ranging from 0.125 to 1.2 based on Transaction Type Functions. See the reference (Tables 3 and 4) on how to set MIR

Maintenance Value Adjustment Factors (MVAF)
The 14 factors are reduced to 4; see reference for details

Functional Similarity
Multiply by .75 for cut and paste type changes
Performance Sensitivity
Multiply by 1.15 if function access large database table.
Complex Function
Evaluation of complexity. Multiply by factors of 1.5 or 3.0 for added functions. Add function points of 3 or 12 for chnaged functions.
Standard Functionality Multiply by 0.5 if language or library provides the required change.

Final Equation
A separate MVAF factor is applied to each individual function transaction type.
This is a lot of detail but keep the objective in mind; it's worth the trouble.

AFP _maintenance =

UFP _{deleted function types} × VAF _pre) × 0.2 +
SUM (UFP _additions × MVAF_{functional similarity} × MVAF_{performance sensitivity} × MVAF_{complex function} × MVAF_{standard functionality}) +
SUM (UFP _changes × MIR _changes × MVAF_{funct. similarity} × MVAF_{perf. sensitivity} × MVAF_{standard functionality} + MVAF_{complex function})

Notes:

SUM => a separate calculation is done for each transaction type.
additions and changes apply only to transaction function types.
MIR _changes is determined by Table 4.
MVAF is the variation on the VAF (4 factors only) described at the end of section 2.

Adapted from Dan Stearns
Last updated 4/12/07