Professional Documents
Culture Documents
Eduardo Miranda
ESCOM
April, 2001
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
More projects have gone awry for lack
of calendar time than for any other
reasons combined.
F. Brook, 1982
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
The Standish Group Report
l The Standish Group research shows a staggering 31.1% of projects
will be canceled before they ever get completed. Further results
indicate 52.7% of projects will cost 189% of their original estimates.
l Resolution Type 1, or project success: The project is completed on-time and on-
budget, with all features and functions as initially specified.
l Resolution Type 2, or project challenged: The project is completed and operational
but over-budget, over the time estimate, and offers fewer features and functions than
originally specified.
l Resolution Type 3, or project impaired: The project is canceled at some point during
the development cycle.
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Many of these projects should have
never started under the conditions
they did
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
The House
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Joe wants a house, ...
210 people
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Epilog
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
The Software
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Bill calls you to his office and says…
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
We won’t be able
to make it.
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Epilog
l …
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
The solution
l Decision rules
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Visualizing software projects
Probability of
being late
Internal
Interfaces
Staff
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Number of people necessary
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
How do programmers spend their time?
Job
Communications
32%
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
How do projects spend their time?
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Interfaces
1400
1200
Comunication Paths
1000
800
600
400
200
0
5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
Number of Subsystems
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Variances on the ISBG data set
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Project Profile - Visualizing the project
complexity
60
50
40
30
20
10
Prob. of being late
0 Interfaces
6 Staff
12
18
24
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Another view of the same profile
6 months
schedule
65 %
Probability of
60 %
being late
30
55 % 24 months
Internal
schedule
5 20 interfaces
10
15
Staff
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Project Profiles Table
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Decision rules
Can change
Criticality
technical
Internal On-time of
People approach or Recommendation
Interfaces Probability Delivery
increase
Date
productivity?
Re-evaluate. Be suspicious of
Yes breakthroughs. Very few
- Have qualified resources - Able to manage improvements pay-off the first time.
High
- Able to build team - Able to test
- Able to manage team - Able to integrate 35% < p <= 50% - Renegotiate schedule
No
- Enough leaders to co- - Renegotiate scope
ordinate
No Go ahead
- Renegotiate schedule
<= 35%
- Renegotiate scope
No to any of the - Renegotiate schedule
above - Renegotiate scope
- Renegotiate schedule
No to any of the above
- Renegotiate scope
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Estimating people, interfaces and
probabilities
l Credible
l Minimum information
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Number of people
ProjectStaff = NominalHeadcount + IntraTeamAdj + InterTeamAdj
ApplicationSize
NominalHeadcount =
Productivity * Schedule
NominalHeadCount
TeamSize =
NoSubsystems
TeamSize * (TeamSize − 1)
IntraTeamAdj = * Overhead
2
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Estimating the number of communication paths
Partitioned
Intra Team Communications
Communication
Model
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Number of interactions / internal interfaces
TeamSize * (TeamSize − 1)
IntraTeamC omm =
2
NoSubsyste ms * ( NoSubsyste ms − 1)
InterTeamC omm =
2
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Sources of schedule variations
Number of un-planned
iterations
+
Activity length and +
variance
- Schedule
Activity familiarity
Uncertainty
+
Performance uncertainty
+
-
Unknown unknowns
Uncertainty reduction
actions
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Schedule uncertainty
Project 1
On-time
latest
completion
probabilities
Project 2 date
Prob. of not
Risk Factor completing the
task on-time (%)
Initial uncertainty value 5
New product development + [2.5, 5]
New technology / process being used + [2.5, 5]
If applicable
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
On-time probability
( Latest − Planned + 10 %) 2
OntimeProb ability = 1 −
( Latest − Earliest ) * ( Latest − Planned )
Earliest = Planned − σ Project
σ Project
= σ 2
Design
+σ 2
Developmen t
+σ 2
Integratio n
σ Design
= DesignLead Time * (Uncertaint y + NoSubsyste ms * .01)
σ Developmen t
= Developmen tLeadTime * (Uncertaint y + ProjectSta ff * .01)
σ Integratio n
= Integratio nLeadTime * (Uncertaint y + InternalIn terfaces * .01)
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Project Profiles Table
HLR SCP
HLR
200~400%
100%
SCP
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001
Concluding remarks
Ericsson Research Canada © 2001, All rights reserved Eduardo Miranda, ESCOM, April 2001