Forming a Winning Programming Team

1
Forming a Winning Programming Team

• Dr. Sallie Henry
• Computer Science Department
• Virginia Tech
• henry_at_cs.vt.edu

2
Necessary Conditions for a Winning Team

• A dedicated coach (willing to spend at least 10
  hours a week)
• Dedicated team members (even if they wont make
  the team THIS year)
• Support from either the Department, the local ACM
  or external source
• Do NOT expect miracles the first year

3
Step 1

• Hold Local contest
• Individuals are competing against each other
• Motivate the students to participate (prizes,
  food, etc.)
• 4-6 problems (of different types)
• 3 hours

4
Step 2

• Selections of team(s)
• Choose at least 6 people
• Based on
• Who wins the local
• Which problems are tackled first
• Year in school (freshman are great)

5
Step 3

• Schedule at least 4 mini-practices
• Each 2 hours
• Have 2 problems for each session
• Put people in teams of 3 (1 machine)

6
Step 4

• Form teams based on the following
• Strengths of team members (you dont want 3
  people with the same background)
• I like a Mathematics person and an Engineer
• Look at Personalities (see who can work well with
  others)

7
Step 5

• Have one 5 hour practice each week
• Compete like a real competition
• Use last years regional problems
• Feed the team during practice (get someone to pay
  for food.
• Have students estimate how long each problem will
  take

8
Step 5 (cont)

• Have another student on the team check the
  results prior to a submission
• Take away egos
• Show them how bad their estimates are
• Attendance is mandatory

9
Step 6

• Give homework
• Have EACH team member write a solution to each
  practice problem
• Collect each problem and source listing in a
  binder

10
Step 7

• The next day
• Have one hour session
• Discuss how to solve the problems
• Discuss the problem selection process
• Have members share ideas

11
Step 8

• Enter at least 2 teams in the regional contest
  (more if you can)
• Make it clear which team is expected to win (make
  one team the strongest)
• The other teams are JV (in training for next
  years contest)
• Early bedtime the night before the region

12
Step 9

• WIN THE REGION

13
Step 10

• Practices after Christmas
• Each week, one 5 hour practice
• The JV teams must attend
• Use prior finals problems
• Homework continues and the binders are getting
  bigger

14
Step 11

• Trip to Library
• Find formula books, algorithm books, etc.
• Assign readings for each student
• Have the team understand the KNOW the information
  in the books

15
Step 12

• Get team ready for finals
• Buy the team shirts with school name on it so
  that they can look like a TEAM during finals.

16
Step 13

• HAVE FUN
• Meet people (your peers) from other institutions
• Look like a team!!
• Take the team out for a nice dinner the night
  before Finals.
• Early bedtime

17
Step 14

• Recruit MONEY
• To by food (breakfast, lunch and/or dinner)
  during practices
• Buy necessary books for the finals
• Buy team shirts
• Pay for the NICE dinner

18
Step 15

• How to get money
• Department
• Local ACM
• Organizations who hire your students

19
QUESTIONS
20
Software DevelopmentTeams Evaluation

• Sallie Henry
• Virginia Tech
• Henry_at_cs.vt.edu

21
Keirsey scale profiles
22
Software Engineering Team Models

• Individual Programmer Mills
• Chief Programmer Mills
• Surgical Team Brooks
• Egoless Team - Weinberg
• Extreme Programming - Beck

23
Belbins Team Roles

• History
• Developed by Meredith Belbin in 1981 at Henley,
  England after nine years of study
• Designed to define and predict success of
  management teams
• Widely used in Europe
• Applied successfully, but widely criticized.
• Measured with the Self Perception Inventory
  (SPI)

24
Belbins Team Roles
25
Belbins Team Roles
26
Belbins Team Roles

• For each individual, there may be multiple roles
• Belbin Primary and Secondary roles
• If primary role is already filled, may gravitate
  toward secondary role
• Strength of roles.

27
The Self-Perception Inventory (SPI)

• Appeared first in Belbins 1981 Book
• Criticisms
• Computer Science Validity
• Defenses
• Positive Field Results
• Confirmed Construct Validity
• Significant Experimental Results
• Use in Industry

28
Previous Results

• Shaper / Leadership
• One is better than two or none.
• Plant
• Presence is better than absence
• Monitor Evaluator
• No conclusive results
• flawed experiment ?

29
The Company Worker

• The implementer
• Theory suggests that this role may affect success
• But this premise was not backed up experimentally

30
Theoretic mapping of MBTI to Belbin