The goal of Data and Analysis Center for Software (DACS) Software & Systems Cost and Performance Analysis Toolkit (S2CPAT) is to capture and analyze software and software engineering data from completed software projects that can be used to improve a) the quality of software –intensive systems and b) the ability to predict the development of software –intensive systems with respect to effort and schedule.
The toolkit allows users to search for similar software projects and use the data to support:
- Rough order of magnitude estimates1 for software development effort and schedule
- Project planning and management: life cycle model information, key risks, lessons learned, templates, estimation heuristics
- Software engineering research.
Figure 1. Overview Statistics
Overview of Current Toolkit Capabilities
Currently, the Software & Systems Cost and Performance Analysis Toolkit can display statistics about all of the software projects in the database as well as allow users to specify search criteria to select software projects of interest based on project size and application domain. The current version of the toolkit provides some standard statistical analyses of the selected projects.
In addition, there is a data submission capability that allows users to submit additional projects to the toolkit so that the toolkit can grow and better reflect changing software development approaches, languages, and tools.
Overview of Statistical Analysis Capabilities
Before searching for information about specific projects in the toolkit, the user can first review general information about the toolkit contents related to distribution of projects by software size and effort, application domain, software operating environment, and development languages, as shown in Figure 1. The upper left chart in Figure 1 shows an example scatter chart with software size on the y-axis and effort (labor hours) on the x-axis. The upper right chart in Figure 1 shows a distribution of the toolkit projects across application domains (note that the current domains primarily reflect DoD types of projects). The two lower charts illustrated project distributions by operating environment (primarily military environments/platforms) and development language.