CohnFord03

Mike Cohn & Doris Ford
Introducing an agile Process to an Organization
IEEE Computer Magazine V36n6(Jun 2003)pp74-78
=EXPERIENCE AGILE PROCESS PEOPLE RISKS Scrum programmer:
1. initially need a balance of scepticism and enthusiasm. Need top talent.
2. Can initially classify Scrum sprints as "prototyping","requirements capture, ... "stabilization".
3. Do not use a distributed development team initially.... especially after a merger. culture clash.
4. Some initially see agile as micromanagment.
Testers: don't let them program! If they want to then hire them as programmers! Testing is up front and managers get involved.
Manager problems: promise customers what? What progress now? How does it effect xyz? When is the end of the project?
Involve Human Resources! Performance reviews without predicted deliverables?

LippertEtal03

Martin Lippert & Petra Beck-Pechau & Jorn Koch & Andreas Kornstadt & Stefan Roock & Axel Schmolitzky & Henning Wolf & Heinz Zullighoven
developing complex Projects using XP with Extensions
IEEE Computer Magazine V36n6(Jun 2003)pp67-73
=EXPERIENCE AGILE XP one-size IT-workplace-solutions

Barry Boehm & Richard Turner
Using Risk to Balance Agile and Plan-Driven Methods
IEEE Computer Magazine V36n6(Jun 2003)pp57-66
=ADVERT one-size situation risk process agile vs planned anchor mbase rup xp spiral modules
Sidebar pp58-59: defines home grounds for agile and plan-based processes plus five critical factors: size, criticality, dynamism, personnel, and culture. Kiviat style polar chart.
sidebar p60:misquotes Cockburn03. 3 levels of skill.
figure 1 p60. idea: if situation does not fit either agile or planned then architect the project into agile and planned parts.
p63: can focus testing on the high-risk parts.
fig 4 p65: 3 kinds of risk: environmental, agile, plan-driven. ranges: minimal..showstopper.

Craig Larman & Victor R Basili
Iterative and Incremental development: A Brief History
IEEE Computer Magazine V36n6(Jun 2003)pp47-56
=HISTORY AGILE METHODS IID evolution1950-2002 non-sequential processes
IID::="iterative and Incremental Development" -- projects that avoided a single-pass sequential document-driven gated-step approach.
Projects/People include 1950s:X-15, Mercury, Los Angeles MIS, IBM, Royce70 (!), Mills, IBM FSD, TRW, LAMPS, Gilb76, Trident, space shuttle, SDC, 1980s: weinberg, McCraken JacksonM, Swartout Balzer, Booch, Boehm Spiral, Brooks86, ParnasClements86, CCPDS-R, Cleanroom, DOD-Std-2167A(!), Gilb88, 1990s: Mil-Std-498 replaces 2167A, Scrum, Shashimi, RAD, DSDM, CAATS, RUP, XP, FDD, CHAOS, 2000: DoD 5000.2, Agile Alliance, ...
Much evidence that IID works better.

Laurie Williams & Alistair Cockburn (eds)
Agile software development: It's about feedback and change
IEEE Computer Magazine V36n6(Jun 2003)pp39-43
=HISTORY AGILE METHODS 1990-2003 DSDM XP Scrum Crystal ASD FDD CMM
Introduces special edition V36n6(Jun 2003)pp39-85
claims that engineers distinguish defined from empirical processes. A defined process can run to completion and produce predicted results automatically. Other processes need short "inspect-adapt" feedback cycles.

Christel Baier & Boudewijn Haverkort & Holger Hermans & Joost-Pieter Katoen
Model-checking algorithms for continuous-time Markov Chains
IEEE Trans Software Engineering V29n6(Jun 2003)pp524-541
=THEORY MATHEMATICS MODEL CHECKING CTMC CSL RELIABILITY CTL
CSL:="Continuous Stochastic Logic". Steady state + path probabilities in time intervals. Modes P[event] <=> p. Nested. Generalizes CTL.
CTMC::="continuous time Markov chains". P(i -> j fires after t) = c*exp( -rate*t) iff it is first.

Robert C Martin
A Better Solution
Software Development Magazine V11n7(Jul 2003)pp53-55
=DIALOG OO CODE NET Java PEOPLE "Micah"
Previous [Martin03e]
Don't send a sequence of strings. Package up a serializable object and send that!