1 - SCOPE reduces the Risk of Projects being Underestimated

Background

Software functionality has multi-dimensional relationships between processes and the data they access. It is the complexity of these logical relationships that is measured and needs to be recorded when assigning function points.

Microsoft Excel Counting

When using a spreadsheet the FP Analyst needs to interpret then transform the physical software into a 3 dimensional logical model then reduce it to a single dimensional list every time they count a new project for an application. It is this interpretation - transformation activity that requires the skill of the Analyst, and it typically consumes around 90% of the counting effort to find, and correctly assess the function. It is also the part of function point counting that is most open to variance since each Analyst will approach it with different information, different experience, different capability and a different modelling strategy. It is this interpretation step that introduces the variations seen by organisations in their project size results typically around 25% when counters use spreadsheets for recording development counts and up to 35% for enhancement counts. Counts are typically undercounted by this amount; it is unusual in our experience to see counts over counted unless the counter has just counted the physical software artefacts such as screens, which they often resort to when they are inexperienced or just too time constrained to find the logical functions.

SCOPE Counting

SCOPEsoftware provides the capability for the FP Analyst to transform the softwares physical model into a 3 dimensional logical model that is stored in SCOPE. All the relationships between processes and data groups can be mapped and recorded down to every field used by a process. Once this structure is recorded, validated and approved for a project then it provides an ongoing baseline framework that can be built on by each subsequent project count so as to ensure ongoing consistency and accuracy of counts. The model only needs to be built once! Inter-count consistency in SCOPE is found to have <10% variance between counters using this methodology of building each count on the previously defined functional model.