Dr. Michael F.

Siok, PE, ESEP

UT Arlington
Computer Science and Engineering
• This slide in work - addressing the questions
1. Why do I have to learn about UML - no one uses it!
2. Why is software engineering different than programming - I'm a very
good programmer
3. What are some of the top challenges in software engineering today?
4. What are some of the latest trends in fortware development?
• https://saturnnetwork.wordpress.com/2010/10/22/five-reasons-
developers-dont-use-uml-and-six-reasons-to-use-it/
• ftp://ftp.cordis.europa.eu/pub/ist/docs/directorate_d/st-ds/softeng.pdf
• http://www0.cs.ucl.ac.uk/staff/A.Finkelstein/talks/10openchall.pdf
• Programming relies mostly • Software Engineering is
on brute force and the learning from other's
strength of the individual failures to avoid the most
programmer (with little common pitfalls
emphasis on learning major
obstacles along the way)
  Famous quotes from well respected Software Engineers
 “The way we build software is in the hunter gatherer
stage.”
 “Cave art. It’s Primitive. We’re computer scientists, but
there’s no science here at all.”
 “If you bought a car with 5,000 defects you’d be very
upset.”

 Douglas Coupland’s “Microserfs”

 A hilarious but frighteningly real look at life in the software
fast lane
 “Shipping hell continued today. Grind, grind, grind. We’ll
never make it. Have I said that already? Why do we
always underestimate our shipping schedules?”
 “In at 9:30 am; out at 11:30 pm - Dominos for dinner. And
three diet Cokes.”

How would you answer the question -

“If we’re ‘software engineers’ what makes us an ‘engineer’?”
• Some typical questions asked
– “How long will it take us (you) to make the change?”
– “Why does it take so long to produce our (your) software?”
– “What can we do to increase our (your) productivity?”
– “Why can’t we (you) just short-cut the process and get this into test (or
release)?”

• How can you satisfactorily answer these questions?

– Hint: this is where the term “Engineering” comes into play

• How do we distinguish Science from Engineering?

– The latter is primarily concerned with the production of a structure,
machine, apparatus, manufacturing process, or a product
Ford's line assigned workers to a
specific production task
 Each production task had a station
where specific
 Work standards, tools, and
measures were utilized

 Standardization of work
 This is the key concept
 The lack of Work standards, tools,
and measures is equivalent to
This work is in the public domain
custom building the automobile

The production line revolutionized the industry producing cars 8x as

fast as before – replacing hand-crafted cars with cars developed by
the standardization of work
• Based on the 1,200+ organizations assessed by the Software Engineering
Institute
• 46 percent have no defined processes
• 32 percent have some basic process defined but at a project level only
• 16 percent have all processes defined at a site/organization level
• 4 percent manage projects using quantitative methods
• 2 percent have continuous improvement processes

• How effective is a football or dance team that doesn’t have a play book or a
choreograph?

• How can work be standardized when almost half of the organizations have no
defined software processes?

• For a partial list of extremely costly software failures visit

http://en.wikipedia.org/wiki/List_of_software_bugs
 Software Engineering is
 The application of a systematic, disciplined, quantifiable approach to
the development, operation, and maintenance of software, and the
study of these approaches; that is, the application of engineering to
software – IEEE SWEBOK

 According to the SWEBOK software engineering activities can be

divided into roughly these ten:
Software Requirements Software Configuration Management
Software Design Software Engineering Management
Software Development Software Development Process
Software Testing Software Engineering Tools
Software Maintenance Software Quality
• Recall that my summary definition of Engineering is
– The ability to use data and/or observations to develop models, laws,
or theorems to aid in the construction of a product
• How is the study of the general production line and techniques related to
engineering?

• Industrial Engineering is concerned with the application of science to the

design of processes and of production systems

• This is accomplished by using engineering techniques to study and model

production line
– Rates (production, defect, and efficiency)
– Processes (quality, bottlenecks, cycle time, push/pull systems)
– Tools (efficiencies, steps, work procedures)

• This is the essential quality of software engineering – development of

models, laws, tools, or theorems to aid in the construction of a product in
a systematic, disciplined, quantifiable manner
 Software development is a very complicated endeavor in industry
 “The application of a systematic, disciplined, quantifiable approach to the
development, operation, and maintenance of software…”
 Unlike an assembly line software development is a creative activity, but
shares many similarities to the assembly line
○ Team development is the norm
○ Large complicated systems require piecewise refinement of requirements , design, etc
○ Work moves from one step to another until a complete product is developed
○ At each step specific tools, techniques and measures are employed to ensure quality –
at each step the work product is transformed
Requirements Design Code Test

Idea Product

 Software Engineering is both a development (production) process and a

creative process
Example Production Line
= + +

Step 1 Step 2 Step 3

What happens if I have a new person Blow glass Create Install filament
blowing the glass bulbs? filament and base

• How does this affect the later steps?

• Where does the assembly error appear
to come from?

• Quality Control vs. Quality Assurance

Tools and procedures help to ensure a product can be produced.

Standards and measures help to ensure it can be efficiently produced.
The “Best” Production Lines have all four:
Tools, Procedures, Standards, and Measures
 The Software Development “Lifecycle”

Requirements Design Code Test

Activities
“stations”
Software Preliminary Code Integration Tests
Requirements Design
Products they
Verification Tests
Interface Detailed Unit Test produce
Requirements Design

UML UML IDE Test Tool

Requirements
Traceability
… Unit Test … Tools used
Tool Tool

Software Preliminary Coding Integration Test

Requirements Design Work Standard Work Standard
Work Standard Work Standard Work
Interface Detailed Unit Testing Verification Test Procedures
Requirements Design Work Standard Work Standard
Work Standard SPE Defect Density - Project A
U Chart (UBAR=0.016), Fagan Inspection, PDS, Lines of Text
Work Standard Fagan Inspection Effectiveness - Project A
X Chart (N=2,XBAR=85.412,MRBAR=7.333), Fagan Inspection, COD, Lines of Code
SPE Defect Density - Project A
U Chart (UBAR=0.004), Fagan Inspection, COD, Lines of Code
SPE Rate - Project A
X Chart (N=2,XBAR=46.586,MRBAR=47.498), Fagan Inspection, PDS, Lines of Text

0.14 120 0.06 800

700
0.12
100 0.05

600
0.1
Major Save s / Lines of Code

Lines of Text / SPE Manhour

Major Save s / Lines of Text

80 0.04
500
Effectiveness (0-100)

0.08

Measures
60 0.03 400

0.06
300
40 0.02
0.04
200

0.01
0.02 20
100

0 0 0
0
Oct-98

Oct-98

Nov-98

Oct-99

Oct-99

Oct-99

Nov-99

Nov-99

Nov-99

Nov-99

Nov-99

Oct-98

Oct-98

Nov-98

Oct-99

Oct-99

Oct-99

Nov-99
Jun-98

Jun-98

Jul-98

Aug-98

Aug-98

Aug-98

Aug-98

Dec-98

Jan-99

Jan-99

Jan-99

Feb-99

Feb-99

Jul-99

Aug-99

Aug-99

Aug-99

Aug-99

Aug-99

Dec-99

Jun-98

Jun-98

Jul-98

Aug-98

Aug-98

Aug-98

Aug-98

Dec-98

Jan-99

Jan-99

Jan-99

Feb-99

Feb-99
Apr-98

Apr-98

May-98

Sep-98

Sep-98

May-99

Apr-99

Apr-99

Apr-99

May-99

May-99

May-99

May-99

May-99

May-99

May-99

Apr-98

Apr-98

May-98

Sep-98

Sep-98

May-99
Mar-98

Mar-99

Mar-99

Mar-99

Mar-99

Mar-99

Mar-99

Mar-98

Mar-99

Mar-99

Mar-99

Mar-99
Nov-99

Nov-99

Nov-99

Nov-99

Nov-99
Jul-99

Aug-99

Aug-99

Aug-99

Aug-99

Aug-99
Apr-99

Apr-99

Apr-99

May-99

May-99

May-99

May-99

May-99

May-99

May-99
Mar-99

Mar-99

SPE Defect Density UCL3 UCL2 UCL1 UBAR LCL1 LCL2 LCL3 Fagan Inspection Effectiveness
Source: UCL326, 2001
Automet as of Mar UCL2 UCL1
8:49 AM Baseline: XBAR JSF Code
Stored: LCL1 FIE LCL2 LCL3 SPE Defect Density UCL3 UCL2 UCL1 UBAR LCL1 LCL2 LCL3 SPE Rate UCL3 UCL2 UCL1 XBAR LCL1 LCL2 LCL3
 Cost Goal

Requirements Design Implement Testing

Escaping Escaping Escaping

Defects Defects Defects
Defect Defect Defect
Removal Removal Removal
• Need to understand how many errors we will have and the re-work cost
associated with them
• Need to remove as many as the budget/schedule will allow and the
delivered product can tolerate
•We will develop and use life-cycles, processes, and methodologies that
deliver the best approach that meet cost, schedule and quality needs
• A software development process transforms the initial system concept
into the operational system running in the target environment.
• It identifies the business needs, conducts a feasibility study, and
formulates the requirements or capabilities that the system must deliver.
• It also designs, implements, tests, and deploys the system to the target
environment.
• The biggest question we will address this semester is "how do I develop a
software product from a customer concept into a working product?"

Copyright © The McGraw-Hill

14
Companies, Inc.
• Should I just slam out some code? How well will that work?

• How do I communicate with the customer and capture the essential thigs
they want in the software?

• How do I decide what the software architecture should look like?

• How do I model user interaction?

• How do I design user interfaces?

• What are design patterns and how should I use them?

Student comments
• Newer discipline
• Easier to revisions
• Innovation/pace of change
• Not physical – more ways it can break
• Laws underneath are more complex
• Can have many purposes, and can change
• Not as much time spent testing
• Not required to be as robust
• Management difficult because hard to measure quality/intangible
• My answers:
• Software is designed, not manufactured
• Little re-use achieved in practice
• Software needs to be practiced and managed like an Engineering
discipline
• Software is malleable
• Can apply to huge variety of problems
• Software doesn’t wear out
• All problems are "designed in"
Software quality assurance (SQA) ensures that
• the development activities are performed properly, and
• the software artifacts produced by the development activities meet
the software requirements and desired quality standards.

Copyright © The McGraw-Hill

18
Companies, Inc.
• Software project management oversees the control
and administration of the development and SQA
activities.
• Project management activities include
– effort estimation
– project planning and scheduling
– risk management
– project administration, and
– others.
These activities ensure that the software system is
delivered on time and within budget.

Copyright © The McGraw-Hill

19
Companies, Inc.
• Object-oriented software engineering (OOSE) is a specialization of
software engineering.

• The object-oriented paradigm views the world and systems as

consisting of objects that relate and interact with each other – my
takeaways on OOSE follow
– It provides the ability to relate the software system being
developed back to the customer in their terms
– OOSE allows the customer to conceptualize users, roles, classes
(objects), and top level interactions from their point of view – it
eliminates a huge conceptual gap
– It builds a conceptual model for the developer of how the
customer’s world operates

• OOSE encompasses:
– OO processes
– OO methodologies
– OO modeling languages
– OO tools

Copyright © The McGraw-Hill

20
Companies, Inc.
• Software engineering aims to significantly improve
software productivity and software quality while
reducing software costs and time to market.

• Software engineering consists of three tracks of

interacting life cycle activities:
– software development process
– software quality assurance, and
– software project management

• Object-oriented (OO) software engineering is a

specialization of software engineering. It helps to
effectively bridge the communication gap between
developer(s) and customer(s)
Copyright © The McGraw-Hill
21
Companies, Inc.
• What are the focuses of computer science and software engineering,
respectively?
• What are the educational objectives of computer science and software
engineering, respectively?
• Some authors say that software engineering is “programming in the large.”
What does this mean?
• What is the relationship between software engineering and computer
science? Can you have one without the other?

Copyright © The McGraw-Hill

22
Companies, Inc.
• To help you be able to develop software - beginning with a customer idea
and ending with the development of a successful product

• To help you apply engineering discipline to the software process through

1. Step-wise refinement of techniques that apply proven principles to
the development of software
2. To help you become very familiar with the latest UML and software
design techniques
3. To provide you with a full-life cycle understanding of software
development
4. To give you some opportunity to work on a software team and
develop an Android product
• Technical content
– Learn the Software Engineering approach using UML
– Develop an Android application for demonstration at semester end
– Deliverables tied to your product

• Management
– Teamwork
– Presentations
– Project Planning

• Experience
– Hands on work with SE processes
– Experience developing Android applications