Cleanroom Software Engineering

1
Cleanroom Software Engineering
Software Testing and Verification Lecture 25

• Prepared by
• Stephen M. Thebaut, Ph.D.
• University of Florida

2
Required Reading and Additional Reference

• Required Reading
• Linger, Cleanroom Software Engineering for
  Zero-Defect Software, Proceedings, 15th Int.
  Conf. on Soft. Eng. (1993), IEEE Computer Society
  Press, pp. 2-13.
• Additional relevant reference
• Linger, Trammell, Cleanroom Software Engineering
  Reference Model. CMU/SEI-96-TR-022, Software
  Engineering Institute, 1996.

3
Cleanroom SE Philosophy

• Cleanroom Software Engineering is a software
  development philosophy.
• First introduced in the 80s within IBM by Harlan
  Mills.
• Based on the notion that defects in software
  should be avoided rather than detected and
  repaired.
• Software development should not be viewed as a
  trial-and-error undertaking.

4
Cleanroom SE Philosophy (contd)

• In traditional software development, errors
  were regarded as inevitable. Programmers were
  urged to get software into execution quickly, and
  techniques for error removal were widely
  encouraged. The sooner the software could be
  written, the sooner debugging could begin.

5
Cleanroom SE Philosophy (contd)

• Today, debugging is understood to be the most
  error-prone process in software development,
  leading to right in the small, wrong in the
  large programs...

6
Characteristics

• Team-oriented
• The functional specification is created by the
  development team, or by a separate specification
  team for large projects, and the usage
  specification is created by the certification
  team.
• Object-based box structure specification and
  design
• Stepwise refinement

7
Characteristics (contd)

• Uses function-theoretic correctness verification
  components are not executed or
  developer-tested!
• Team correctness verification takes the place
  of unit testing and debugging, and software
  enters system testing directly, with no execution
  by the development team...no private debugging
  (is) permitted.

8
Characteristics (contd)

• Statistical usage testing (of integrated
  increments) is undertaken for quality
  certification (statistical quality control).
• The certification (test) team is responsible
  for...certifying the quality of software with
  respect to its specification. Certification is
  carried out by statistical usage testing that
  produces objective assessments of product
  quality.

9
Characteristics (contd)

• Incremental development
• Management planning and control...is based on
  developing and certifying a pipeline of software
  increments that accumulate to the final product.
• Structured programming

10
Impact on Development Cycle

• Experienced...teams...can achieve substantially
  reduced product development cycles. The precision
  of Cleanroom development eliminates rework and
  results in dramatically reduced time for
  certification testing compared to traditional
  methods. And Cleanroom teams are not hostage to
  error correction following product release.

11
Box Structure Specification and Design

• Incorporates black box (external behavior), state
  box (retained data), and clear box (processing)
  forms.
• Transition Functions
• Black box (S, SH -gt R)
• State box (S, OS) -gt (R, NS)
• Clear box (S, OS) -gt (R, NS) by procedure
  (intended function)
• Intended functions are refined into control
  structures (programs)

12
Verification

• Development teams employ mental proofs of
  correctness in team reviews
• Every correctness condition of every control
  structure is verified every team member must
  agree that each condition is correct.

13
Quality Certification

• Based on statistical quality control in
  manufacturing
• Process (statistical usage testing)
• sample population of user executions based on
  expected frequency (stratified random sampling)
  operational profile
• measure quality by determining if executions are
  correct
• extrapolate to the population of possible
  executions (statistical inference)
• if quality is inadequate, identify and correct
  flaws in development process

(contd)
14
Quality Certification (contd)

• Alternate distributions can be defined for
  low-probability, high-consequence functions.
• Errors tend to be found in failure-rate order on
  average (coverage testing is not biased to find
  errors in any particular rate order).

15
Cleanroom Software Engineering
Software Testing and Verification Lecture 25

• Prepared by
• Stephen M. Thebaut, Ph.D.
• University of Florida