Towards Estimating Schedule for COSYSMO:�Workshop III

2024 BCSSE COCOMO Forum

1

11/14

BCSSE COCOMO Forum

The Aerospace Corporation & Online

November 14, 2024 at 08:45 PST

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Overall Agenda

• Questionnaire
• The Issue
• Summary of some previous presentations and discussion
• The Valerdi 2009 paper
• The Valerdi & Wheaton “Staffing Profile Chart”
• Is there a “Consult” phase?
• Jim tries a theory
• Brad makes a suggestion

​

​

​

​

2024 BCSSE COCOMO Forum

2

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

A Questionnaire for You

Soliciting your opinions and further cooperation in working on a schedule model for COSYSMO

• The questionnaire is an Excel file asking for your opinions on a few multiple-choice questions on this topic
• After you fill it out, please E-mail it to me per instructions in the file.
• Thanx

​

2024 BCSSE COCOMO Forum

3

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

The Issue at Hand

COSYSMO is a family of systems engineering effort estimating models

• The most recent of these is my COSYSMO 3.0 model [1]
• However, there does not seem to be any schedule estimating model in the literature
• For comparison, COCOMO II [2] does include a software schedule estimating model

This presentation (and previous versions of it [3]) are my attempt to try to develop such a schedule estimating model

• A scope note: The issue is models for systems engineering, not models for a project as a whole

​

​

​

​

2024 BCSSE COCOMO Forum

4

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

History of COSYSMO Models

2024 BCSSE COCOMO Forum

5

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

On the Paper by Valerdi et al - 1

• Mauricio Pena pointed me to Ricardo Valerdi’s 2009 paper
• “Approaches to Calculating Systems Engineering Schedule in Parametric Cost Models” [4]
• I spent a fair amount of time with the paper, but had a lot of difficulty:
• The point of the paper seemed to be that COCOMO-type estimating equations could be derived from management theory
• At least part of the paper applied economic principles that seem incorrect in this context
• Doubling the number of engineers on a project will cut the time in half
• Parkinson’s Law (“Work expands to fill the time available for it”)
• Brooks’ Law (“Adding staff to a late project makes it later”)

​

​

​

​

​

​

​

2024 BCSSE COCOMO Forum

6

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

On the Paper by Valerdi et al - 2

• (Continuing)
• After several management-theory equations, COCOMO-type estimating models for size and duration were derived
• The Next Steps section includes this:
• “A logical next step is to integrate these approaches into the COSYSMO model to allow users to estimate systems engineering effort and schedule.”
• I am unaware of any work on “the logical next step”

​

​

​

​

​

​

2024 BCSSE COCOMO Forum

7

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Valerdi & Wheaton’s “Staffing Profile Chart”

• [5] includes this “staffing profile chart” as its Figure 2 “Systems Engineering Effort Profile”:

​

​

​

​

​

​

​

​

• The horizontal axis must be time, divided into Life Cycle Process phases
• The chart therefore may be interpreted as providing duration information

2024 BCSSE COCOMO Forum

8

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Further Analyze the Staffing Profile Chart?

• Two kinds of information displayed in the chart might perhaps be interpreted to provide duration information:
• The width of each of the four ISO/IEC 15288 phases (duration information);
• The shapes (height = staffing level, vs time) of each phase for each ANSI/EIA 632 fundamental process (staffing level over time information)
• At best, the chart by itself would only provide relative information (e.g., relative staffing levels)

​

​

​

2024 BCSSE COCOMO Forum

9

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Further Discussion with Wheaton

• At this conference last year [3], Marilee made a couple of clarifying comments:
• Remembering development of Staffing Profile Chart—some aspects were more notional
• Possibly Staffing Profile Chart is not so representative of actual projects
• My conclusion: No specific schedule information can be derived from this chart

​

​

​

2024 BCSSE COCOMO Forum

10

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Jim’s Notional Schedule Diagram [3]

• System Engineering effort for new development is expended in a mostly bimodal distribution, where during a Consult phase system engineers merely support other disciplines.

2024 BCSSE COCOMO Forum

11

11/14

Conceptualize; Develop

Operational Test & Evaluation; Transition to Operation

Consult

Time

Headcount

Example issues:

• What duration is needed for Conceptualize + Develop? For OT&E + Transition? (These can be absolute, or %ages.) Alternatively, what average headcount?
• How should effort be divided among the three phases?
• What headcount is needed for Consulting? Zero?

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Jim’s Notional Schedule Diagram (Expanded)

2024 BCSSE COCOMO Forum

12

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Conclusion of “Consult” Discussions from 2023 Forum Workshop

​

​

• A few attendees seemed to feel like there was a Consult phase with reduced Systems Engineering effort
• Though possibly not as depicted above
• But most attendees seemed to believe that such a phenomenon didn’t occur

2024 BCSSE COCOMO Forum

13

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

A Schedule Equation

• Given an effort estimate in person-months, these equations provide a schedule estimate (TDEV):

​

​

​

​

​

​

2024 BCSSE COCOMO Forum

14

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

“Where Have I Seen This Before?”

• Of course, this is the software effort-to-schedule estimation model from COCOMO II [3].
• The COCOMO II scale factors are:
• Precedentedness
• Development flexibility
• Risk resolution
• (Stakeholder) team cohesion
• Process maturity
• The sum of scale factors ranges from 0 to 31.62
• Therefore F ranges from 0.28 to 0.34324.
• Note also that the scale factors play an exponential role in the effort calculation

​

​

​

​

2024 BCSSE COCOMO Forum

15

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Analogous Equation for COSYSMO

• This is the analogous COSYSMO schedule equation

​

​

​

​

• COSYSMO 3.0 scale factors:
• Risk/Opportunity Resolution
• Process Capability
• Requirements Volatility
• Can’t determine X, Y, & Z without calibration to a data set
• Can maybe estimate by analogy?

​

​

​

​

​

2024 BCSSE COCOMO Forum

16

11/14

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction

Bibliography

1. Alstad, J.P., “Development of COSYSMO 3.0: An Extended, Unified Cost Estimating Model for Systems Engineering”, Procedia Computer Science 2019, pp 55-62.
2. B. Boehm, et al, Software Cost Estimating with COCOMO®️ II, Prentice Hall, 2000
3. Alstad, J.P., “Towards Estimating Schedule for COSYSMO:�Workshop II”, Boehm Annual Research Review (BCSSE online), April 2024.
4. Gaffney, John E., Valerdi, Ricardo., Ross, Michael A. “Approaches to Calculating Systems Engineering Schedule in Parametric Cost Models”, CSER 2009, April 2009.
5. R.Valerdi, M.Wheaton, “ANSI/EIA632 As a Standard WBS for COSYSMO,” AIAA 1st Infotech@Aerospace Conference, Arlington, VA, September 2005.

​

​

​

​

​

2024 BCSSE COCOMO Forum

17

11/14

–

Boehm CSSE

Center for Systems and Software Engineering

ESSE: Introduction