UNCLASSIFIED
IDA Lai No. HO 92-041141
DEFINITION$
ID ob~AeOft$ e foltoer togevaw*Ut to fled the a. vfti of aS wors
Rapofft1401110M ait the Most aa 11t"lt~a' agE iMust Ut ~lty 9*&U#4vVtfisituft ICA 10 0AfS~It they Normally eso"" ("Viott of maow PtO##cti g tsI "S
1.sawtr 6481149 " dirntwn Aflectlag Major protrais, (a) address rtits" of Ut4*t0M eexcet to m Executive 3iach. I the res C Ogftti Tlehe "61wi, or lei am#"l~ It"*# testl 110* sisifianta oetoemtc imoeluaums WA* ftpom are tio ftwtd *o* Ponefta oufs *w#U to suremo tht"rhg quality sod flifeafts, to ft. pobloor, Ibed. In support of this objective, IDA had previously developed a high-level concept of a research and development initiative that would address deficiencies in required SDI software testing technology. The purpose of this paper is to provide SDIO with a set of recommendations and supporting analysis to begin a comprehensive software testing initiative in fiscal year 1993.
IVWith the Missile Defense Act of 1991, Congress appropriated funds for the development and deployment of ballistic missile defense systems capable of protecting the United States, as well as US forces, friends, and allies overseas. Several of the experimental weapon and sensor program 3 funded by the Strategic D)efense Initiative Organization (SDIO) will make the transition to full-scale development in the next few years. All of these systems will rely upon software to function properly.Historically, 50% to 80% of the c~st of software development is spent in testing. There is no reason to expect that testing will play a lesser role in Strategic Defense Initiative (SDI) software development. Clearly, software testing will play a critical role in developing reliable and cost-effective SDI software. However, previous studies by the Institute for Defense Analyses (IDA) have suggested that conventional testing methods sufficient for small-scale sequential software may not be adequate for testing software for strategic defense systems. These methods are largely ad hoc and may not scale up to the levels required by SDI systems. In addition, there are currently no concerted efforts to ensure that existing advanced methods for testing software are used within the SDI program. This paper lays the groundwork for a technical initiative that will develop and deploy software testing technology needed to ensure the development of reliable and costeffective software for SDI. Three objectives for this initiative have been identified: a. To ensure that all element program offices have consistent approaches to testing SDI software.b. To experiment with new and improved testing techniques and to facilitate their transition into SDI standard practice.c. To strengthen the scientific basis for software testing methods, tools, and metrics related to the SDI software testing domain.A specific approach has been developed for achieving each of these objectives. These approaches involve technology transition projects, software testing experiments, and applied research projects.vii Transition Projects
9The fir...