3 ;;Quote: Cleanroom projects achieve 3.3 errors/KLOC in testing after correctness verification

3 ;;Quote: errors in Cleanroom projects tend to be simple mistakes found by statistical testing, not deep design or interface errors

5 ;;Quote: a Cleanroom project developed 107 KLOCs in 3 increments; 50 people, 2.6 errors/KLOC and 486 LOC/developer/month

7 ;;Quote: box structures avoid the ambiguities and omissions for natural language specifications; detects gaps and misunderstandings early

7 ;;Quote: box structure design: all data and processing in boxes, boxes form a usage hierarchy, black/clear/state boxes

7+;;Quote: black, state, and clear boxes have identical external behavior but increasing internal detail

8 ;;Quote: a black box specifies all visible behavior; i.e., a response for each stimulus and stimulus history, including erroneous histories

8 ;;Quote: record black box specifications by tables of stimulus and history conditions; scale up by behavior classes and specification functions

8 ;;Quote: a state box specification replaces the stimulus history of a black box with old state to new state transitions

8 ;;Quote: a clear box implements a state box with sequence, iteration, and concurrency; may invoke black boxes

8 ;;Quote: a clear box forces cooperation of lower level black boxes; defines a usage hierarchy and box interfaces

9 ;;Quote: a program function transforms input state to output state; a clear box includes its intended function and the intended functions of its refinements

9 ;;Quote: confirm that a program implements an intended function by correctness conditions for sequence, alternation, and iteration

10 ;;Quote: mentally verify clear boxes by verifying a small number of correctness conditions; can check unanimously in a few seconds

12 ;;Quote: traditional coverage testing finds errors in random order, usage testing finds them in failure-rate order, 20x better for MTTF