Mason03

John Mason
Comments Considered Harmful
ACM SIGCSE Bulletin V35n2(Jun 2003)pp120-122
=ESSAY COMMENTS STYLE TECHNICAL EDUCATION
(1) Parody of Dijkstra's letter written as student
(2) Anecdotal experience in industry.
(3) ideas used while teaching.
Proposes "reversed assignments": given larger program with bad style and are asked to modify it.

CalcagnoOHearnBornat03

C Calcagno & P OHearn & R Bornat
Program logic and equivalence in the presence of garbage collection
Theoretical Computer Science V298n3(??? 2003)pp557-581
=THEORY FORMAL LOGIC HOARE Garbage Collection (D.4.2) VERIFICATION
What do we mean when we say that a particular object "exists" on the heap? To preserve Hoare's assignment axiom in the presence of garbage collection, the existence of garbage becomes "virtual".
The wise program prover should use a language with proved automatic garbage collection, and avoid assertions that refer to garbage!
Includes a formal model of garbage collection.

Franck Barbier & Brian Henderson-Sellers & Annig Le Parc-Lacarelle & Jean-Michele Bruel mformalizationof the Whole-Part Relationship in the Unified Modeling Language
IEEE Trans Software Engineering V29n5(May 2003)pp459-470
=PROPOSAL Object-Oriented MODEL STANDARD aggregation composition whole-part bibliography UML OCL MetaModel OML OPEN
careful philosophical analysis of semantics of whole-part. some confused "common logic" and much OCL.
Defines a new dotted diamond whole-part Relationship with specialization of aggregation and composition.
Introduces new stereotypes to indicate OCL constraints on the UML MetaModel.
Will it make it into UML4.0?

Stacy J Prowell & Jesse H Poore
Foundations of Sequence-Based software specification
IEEE Trans Software Engineering V29n5(May 2003)pp417-429
=THEORY SPECIFICATION CLEANROOM BSDM
BSDM::="Box Structure Development Method", based on Harlan Mills and CLEANROOM techniques.
Specifications in form of map spec:#input->(empty | output | undefined).
(above) |- #input partitioned into (#input)/spec. -- Mealy equivalence
Use a table to enumerate elements in #input/spec and tracing them to requirements.
Note: does not refer to Hartmannis and Stearns theory.

F E Emery & E L Trist
Socio-technical Systems
Management Science, Models and techniques, V2(?? 1960)pp83-97
=ESSAY SYSTEMS PEOPLE TEAM
Similar observations in British coal mining and Indian textile mills: a composite group/team oriented organization with complex roles and interaction outperforms a rigid hierarchical organization. Also less rigid form ad less symptoms of stress (illness, accidents, etc).
Managers should manage interactions between teams rather than micro-manage the internal structures of teams.
(dick) |- probably applies to agile vs rigid processes.

Min Xie & Bo Yang
A Study of the Effect of Imperfect Debugging on software development cost
IEEE Trans Software Engineering V29n5(May 2003)pp471-473
=THEORY MINIMAL COST PROBABILISTIC STOCHASTIC ECONOMIC MODEL TESTING
Assumes effectiveness of testing effects cost.
Shows how to find optimal testing effectiveness and time to minimize total cost of testing, removing bugs while testing, and removing bugs in the field.

George Marsaglia
Seeds for Random Number Generator
Commun ACM V46n5(May 2003)pp90-93
=IDEA TECHNICAL RNG
RNG::="Random Number Generator".
|-A single seed RNG generates a (long?) but single cycle of numbers.
If x is followed by y, it always flowed by y.
(above) |- single seed RNGs can generate one out of many possible sequences.
|-Using such a RNG to generate a jury list is probably illegal!
Gives 8 examples (in C) of RNGs that use a buffered of seeds and generate more permutations and have a much longer cycle.
Suggests ways of getting the hundreds or thousands of random seeds.
Does not mention the use of a single seeded RNG to shuffle a buffer of previously generated numbers.