Project Cost Estimating

Software cost estimating for large software projects is far too complex to be performed manually. This observation is supported by the presence of at least 75 comcommercial software cost estimating tools, including such well-known tools as COCOMO II, CostXpert, Knowledge- Plan, PRICE-S, SEER-SEM, SLIM, and the like. Successful projects all use at least one such tool, and usage of two or more is not uncommon. Estimates producedby trained estimating specialists are also noted on many successful large projects, but not on failing projects.

• Software estimating tools (COCOMO II, CostXpert, KnowledgePLAN, PRICE-S, SEER-SEM, SLIM, etc.).
• Formal sizing approaches for major deliverables based on function points.
• Comparison of estimates to historical data from similar projects. • Availability of trained estimating specialists or project managers.
• Inclusion of new and changing requirements in the estimate. • Inclusion of quality estimation as well as schedule and cost estimation.

Failing projects tend to understate the size of the work to be accomplished due to inadequate sizing approaches. Failing projects also omit quality estimates, which are a major omission since excessive defect levels slow down testing to a standstill. Overestimating productivity rates or assuming that productivity on a large system will be equal to productivity on small projects are other common reasons for cost and schedule overruns. The main problem with estimates for projects in the 10,000-function point size range is that they err on the side of excessive optimism. Project planning tools and project estimating tools overlap in functionality, and are usually marketed separately. Normally, the project planning and cost estimating tools pass information back and forth. The software cost estimating tool would be used for overall project sizing, resource estimating, and quality estimating. The project-planning tool would be used for critical path analysis, detailed scheduling, and for work breakdown structures.