The Mythical Man Month

1
The Mythical Man Month

• Frederick P. Brooks, Jr

2
Projects gone wrong for lack of calendar time

• The assumption that every will go well Optimism
• Confusing effort with progress Man-Month
• Schedule process is poorly monitored
• Cant meet deadline? Add manpower.

3
Optimism

• Ideas incomplete and inconsistent
• Implementation
• Interaction
• The assumption that all will go well has a
  probabilistic effect on the schedule.
• The Mind of the Maker, Dorothy Sayers

4
The Man-Month

• Cost varies as the product of the number of men
  and the number of months
• Progress does not
• You cant measure the size of a job with the
  man-month as a unit.
• Only when a task can be partitioned among many
  workers with no communication

5
Graphs

• A task that is perfectly partitionable without
  communication
• An unpartionable task due to sequential
  constraints
• Partitionable task with communication
• Task with complex interrelationships

6
Systems Test

• The part of software development schedule that
  gets maximum affected
• By sequential constraints are
• Component Debugging
• System Test
• The time required depends on number and subtlety
  of the errors occurred.
• Theoretically this number should be zero.

FOR MORE INFO...
7
Rule of Thumb

• Rule of Thumb Scheduling Software
• 1/3 planning
• 1/6 coding
• 1/4 component test early system test
• 1/4 system test, all components in hand
• This differs from Conventional Scheduling

8
Contrast
9
Conventional System Testing

• Failure to provide enough time for system test
  may prove disastrous.
• Since delay comes at the end of schedule no one
  is aware of the schedule trouble till the
  delivery date.
• Delay at this point has several repercussions.
• Financial
• Psychological
• Secondary costs may suffer too.

10
Different Types Systems Testing
11
Estimation

• The urgency of the customer may govern the
  scheduled completion of a task, but it cannot
  govern the actual completion.

12
Gutless Estimating

• It is very difficult to make a vigorous,
  plausible and job risking defense of an estimate
  that is derived by no quantitative method,
  supported by little data, and certified chiefly
  by hunches of the managers.
• Frederick P. Brooks
• Need to develop publicize productivity figures,
  bug incidence figures, estimating rules etc.
• While estimating individual managers will need to
  stiffen their backbones and defend estimates.

SOLUTIONS...
13
Gutless Estimating

• What is it?
• Assurance that a hunch are better than wish
  derived estimates
• Urgency of client causes optimistic estimates
  (dictionary.com)
• Some examples of it
• A cook
• Software engineers
• Two Solutions
• Need to develop and publicize productivity
  figures
• Need to stick to your hunch

14
Regenerative Schedule Disaster

• When a software project is late, add more man
  power
• Add one more man to the team
• Add two more men to the team
• A child takes 9 months no matter how many women
  you get
• Brooks law "Adding manpower to a late software
  project makes it later."
• http//javatroopers.com/Mythical_Man_Month.html

15
How to estimate correctly

• Number of months depends on sequential
  constraints
• Max manpower depends on independent subtasks
• From these two we can use fewer men and more
  months
• Down fall to this is fear of obsolescence

16
Questions and Comments

• Kay Abma
• Amritam Sarcar
• Ruben Garcia