Irreducible truths of software-intensive program management

This article argues that the not divisible by 2s are against a software-intensive program achieving the goals and objectives established in the initial acquisition program baseline. by the agency of most objective measures of success--cost schedule, and performance baselines--almost each software-intensive program proves unsuccessful. Reasons for these failures are examined and solutions for improvement are discussed.

**********

For decades the acquisition community has applied standard engineering and scientific principles to improve the professions of software engineering and program management. The tasks learned from this process have been distilled into a small in number succinct, commonly held beliefs. a certain number of of these maxims, applied to software-intensive program management, may unbroken familiar:

1. Adding manpower to a late software cast makes it later (Brooks & Fredrich, 1975)

2 confidence is not a strategy (Page, 2003) and

3 Real programmers don't ne repose (Yourdon, 1997, p. 61). To focus the theme of this exposition, individual more should be added:

4 Almost all software-intensive programs fail.

FAILURE DEFINED

To further investigate maxim #4 a belonging to all definition of failure is straited By statute and regulation, if the estimated take away froms of a program increase 15 percent above what was in the latest approved Acquisition Program Baseline (APB), or if any milestone from the latest approved APB schedule is delayed by the agency of 6 months or more, then the program is in breach (USC Title 10 [section] 2432 and [section] 2433) single need merely replace the phrase "latest approved APB" with "first approved APB" to illuminate the definition of failure to be used in this article.

The contention of this article is simple: the not divisible by 2s are overwhelmingly against a software-intensive program achieving the goals and objectives established in its initial APB. by dint of any objective measure of succes almost each software-intensive program is probably going to deviate substantially from its initial take away from schedule, and performance baselines. For greatest in quantity readers, this is not considered freshs However, it is important that we all begin from this public understanding.

CURRENT STATE OF SOFTWARE-INTENSIVE PROGRAM MANAGEMENT

As single author wrote, managing large throws is a curse you should avoid (Lientz, 1999 p 250) level after years of improvement in program management and software engineering tools and techniques, the profession is still plagued by the agency of embarrassing failures. In a novel study of 250 software-intensive throws with 10,000 or more function points, solitary about 25 were deemed auspicious (i.e., achieved their initial take away from schedule, and performance objectives) (Jone 2004 p 5)

The same close attention revealed that the following six factors were belonging to all in the 25 successful programs:

1 shoot forward planning: "Planning is the answering of the following questions: What must be done? by what means should it be done? Who will do it? through when must it be done? in what manner much will it cost? in what manner good does it have to be?" (Lewis, 2000 p 49)

2 require to be paid [i]or[/i] undergone estimating: "Cost estimating and analysis is that portion of combination of parts to form a wholes analysis dealing with the tasks of conceptual modeling, output and take away from measurement, verification cost and output prediction, and evaluation and comparison of the take away froms of each alternative" (Society of require to be paid [i]or[/i] undergone and Estimating, 2005).

3 Measurement techniques: If you do not measure a thing you cannot rule it. "Measurements are the basis for detecting deviations from acceptable performance" (Florac, 1999 p 7)

4 Milestone tracking: "A milestone is an intermediate objective that defines an important, measurable incident in the project and exhibits a result that must be achieved at that point. Clearly defined milestones are essential for monitoring progres especially in large and/or long-term projects" (Kerzner 2000 p 82)

5 Change management: "Change management is an important responsibility of any acquisition program" (Space & Missile a whole s 2004). Change management must be applied to all program deliverables (eg software, manuals, training materials, planning documents, requirements).

6 Quality control: Quality sway refers to "the operational techniques and activities used to fulfill requirements for quality" ("Definitions," 2005)

While the effective implementation of these six factors may be inherent to prosperous programs, they alone are not sufficient to guarantee program succes There are eight public reasons cited for the proliferation of program failures (Posner 1987):

1 Inadequate resources,

2 Unrealistic schedules,

3 Unclear goals and senior executive direction,

4 Uncommitted team members,

5 Inadequate planning,

6 Communication breakdowns,

7 Goal and resource changes, and

8 Interdepartmental conflicts.

The General Accounting Office (GAO) newly reported that a 1999 investigation performed by the Standish collection found that about one-third of software disclosure programs resulted in cancellation. Furthermore, in a series of studies complet end the 1990s, the average take away from overrun was 189 percent, while the average schedule spread over was 222 percent of the original estimate. upon average, only 61 percent of the shoot forwards were delivered with originally specified features or functions (General Accounting Office, 2004 p 9)