CSCE 606

Software Engineering

Hank Walker

Based on materials from Armando Fox and David Patterson

© 2013 Armando Fox & David Patterson, all rights reserved
1
 Outline
Class Organization
§1.1 Introduction to Software Engineering
§1.5 Software as a Service
§1.4 Service Oriented Architecture
§1.6 Cloud Computing
§1.7 Beautiful vs. Legacy Code
§1.8 Software Quality Assurance: Testing
§1.9 Productivity: Clarity via Conciseness,
Synthesis, Reuse, & Tools
2
 Course Goals
 Learn Software Engineering Principles by
understanding new challenges,
opportunities, and open problems of SaaS
 Take a SaaS project from conception to
public deployment
 Solve Non-Technical Customer problem
 Server side: Ruby on Rails
 Client side: HTML, CSS, AJAX, JavaScript
 Deploy using cloud computing

3
 Projects for Non-Profits
 Teams of ~6 students build projects for non-
technical customers from nearby nonprofit
organizations, e.g.,
 Humane Society Pet Matchmaker
 Student Dormitory Package Notifier
 Minority VC firm Customer Relationship
Manager (tracks startup proposals)
 Children’s Hospital Nurse Vacation Scheduler
 Teams meet weekly with TA; start forming
this week!
4
 Children’s Hospital
 Screencast of Spring 2012 UCB CS169
project Nurse Scheduler
https://vimeo.com/44837891
 Interview with Nurse Managers
https://vimeo.com/46770083
 Your team will record a video interview and
screencast for your project

5
 Results/Observations
 Non-technical customer feedback
 90% of teams reported customers “happy” or “thrilled”
 50% of customers tried to hire students to continue work
 67% of students intend to maintain app after semester
 Students engaged in process
 Later iteration stories more uniform in complexity & size
 Projects varied in code quantity but rarely quality

7
 Course Organization
 Grading
 20% assignments
 15% quizzes
 15% midterm
 50% project
 Bonus points based on Participation and Altruism (e.g.,
answering Piazza questions)
 Schedule
 TR: lecture 5:30-6:45pm, HRBB 124
 Teams meet TA (and me) by appointment
 Use Piazza (and eCampus)
8
 Prereqs & Course Format
 Prereqs
 Official: CSCE 431 or approval of instructor
 Practical: Know object-oriented programming language
 Format
 Before lecture: do reading
 In lecture: put reading in context
 After lecture: assignments, for hands-on practice
 Required “2-pizza team” (~6 students) project
 Design, develop, deploy to “production ISP” (Heroku)
 Outsiders invited to final project poster session

9
 Programming Homework
 Programming homework done on your own
 Late policy (without good excuse)
 ¾ credit if 1 day late, ½ credit if 2 days late, 0 if later
 Assignments and project will be deployed using cloud
computing and evaluated there
 AWS Cloud9 development environment – sign up for free
account, must give them your CC
 https://aws.amazon.com/cloud9
 http://www.saasbook.info/bookware-vm-instructions
 GitHub: Get a Free Micro Account for Students for use in
this class (not TAMU GitHub)
 http://education.github.com
 We will use many other free open source tools
10
 Honor Code
 The Aggie Honor Code:
“An Aggie does not lie, cheat, or steal or
tolerate those who do.”
 If you wouldn’t want fellow students, parents,
or professors to know about it, then stop it!
 Your instructors join you in pledging to
adhere to this code


11
 Textbook
 http://saasbook.info
 1st edition (v1.2.2)
 Available from Amazon
 Ebook ($10)
 Kindle format
 Kept up-to-date
 Live links
 Print book ($38)

12
 Teaching Assistant
 Vihang Patel
 [email protected]
 Once project starts, hold weekly meetings
with him to discuss project status, problems

13
 Online Resources
 http://www.saasbook.info/students
 Online “get started” tutorials on Ruby in
case you want to start on Ruby now
 Pointers to online HTML/CSS tutorials in
case you want start on Rails now
 Tutorials on other tools (GitHub, c9, etc.)


14
YOUR BRAIN ON COMPUTERS; Hooked on
Gadgets, and Paying a Mental Price
NY Times, June 7, 2010, by Matt Richtel This is your brain on computers.
SAN FRANCISCO -- When one of the most important e- Scientists say juggling e-mail, phone calls and other incoming
mail messages of his life landed in his in-box a few years information can change how people think and behave. They say
ago, Kord Campbell overlooked it. our ability to focus is being undermined by bursts of information.
Not just for a day or two, but 12 days. He finally saw it These play to a primitive impulse to respond to
while sifting through old messages: a big company
wanted to buy his Internet start-up. immediate opportunities and threats. The
''I stood up from my desk and said, 'Oh my God, oh my stimulation provokes excitement -- a dopamine
God, oh my God,' '' Mr. Campbell said. ''It's kind of hard squirt -- that researchers say can be addictive.
to miss an e-mail like that, but I did.''
In its absence, people feel bored.
The message had slipped by him amid an electronic
The resulting distractions can have deadly consequences, as
flood: two computer screens alive with e-mail, instant
when cellphone-wielding drivers and train engineers cause
messages, online chats, a Web browser and the
wrecks. And for millions of people like Mr. Campbell, these
computer code he was writing.
urges can inflict nicks and cuts on creativity and deep thought,
While he managed to salvage the $1.3 million deal after interrupting work and family life.
apologizing to his suitor, Mr. Campbell continues to
While many people say multitasking makes them more
struggle with the effects of the deluge of data. Even after
he unplugs, he craves the stimulation he gets from his Heavy
productive, research shows otherwise.
electronic gadgets. He forgets things like dinner plans, multitaskers actually have more trouble focusing
and he has trouble focusing on his family. His wife,
and shutting out irrelevant information, scientists
Brenda, complains, ''It
seems like he can no
say, and they experience more stress.
longer be fully in the moment.''
And scientists are discovering that even after
the multitasking ends, fractured thinking and
lack of focus persist. In other words, this is also
your brain off computers. 15
 The Rules
(and I really mean it!)

16
05/17/2020
 Peer Instruction Questions
 Increase real-time learning in lecture,
test understanding of concepts vs.
details
 As complete a “segment”
ask multiple choice question
 <1 minute: decide yourself, vote
 <2 minutes: discuss in pairs,
then team vote; flash card to pick answer
 Try to convince partner; learn by teaching

 Mark and save flash cards

1 2 3 4
 Have You Used GitHub?

☐ Yes, I have used GitHub

☐ No, I have not used GitHub

☐ What is GitHub?

☐ I’ve already lost my yellow 3x5 card

18
Which is True About CSCE 606
Schedule?
☐ I can leave for Thanksgiving early since
we won’t have a lecture the week before
☐ I can leave for Christmas early since the
project poster/demo session is optional

☐ Prof. Walker doesn’t mind if I read email

or use Facebook during lecture on my
cell/laptop. He does it too.
☐ You project team needs to hold
weekly meetings with the TA
19
END
20
 Introduction to Software
Engineering

(Engineering Software as a Service §1.1, 1.5)

21
 Ranking Top 200 Jobs (2012)

1. Software Engineer 104. Airline Pilot

28. Civil Engineer 133. Fashion Designer
38. Nurse 137. High School Teacher
40. Physician 163. Police Officer
47. Accountant 173. Flight Attendant
60. Mechanical Engineer 185. Firefighter
73. Electrical Engineer 196. Newspaper Reporter
87. Attorney
InformationWeek 5/15/12. Based on salary, stress levels, hiring outlook,
physical demands, and work environment (www.careercast.com) 22
 If SW Engineering So Popular,
Why So Many SWE Disasters?
 Therac-25 lethal radiation overdose
 1985: Reused SW from machine with HW interlock on
machine without it: SW bug => 3+ died
 Mars Climate Orbiter disintegration
 1999: SW used wrong units (pound-seconds vs.
newton-seconds) => wasted $325M
 FBI Virtual Case File project abandonment
 2005: give up after 5 years of work => wasted $170M
 Ariane 5 rocket explosion
 1996 => wasted $370M

23
 How Can You Avoid Infamy?
 Lessons from 60 years of SW development
 This class will review many approaches in
lecture, listing pros and cons
 Will get hands on experience with one
approach good for Software as a Service

24
END
25
 Software as a Service

(Engineering Software as a Service §1.2)

26
 Software Targets
 Traditional SW: binary code installed and
runs wholly on client device, which users
must upgrade repeatedly
 Must work with many versions of hardware, OS
 New versions must go through extensive
release cycle to ensure compatibility
 Alternative: develop SW that only needs to
work on one HW & OS platform?
 Quicker release cycle, no user upgrades?

27
 Software as a Service: SaaS
 SaaS delivers SW & data as service over
Internet via thin program (e.g., browser)
running on client device
 Search, social networking, video
 Now also SaaS version of traditional SW
 E.g., Microsoft Office 365, TurboTax Online
 Instructors think SaaS is revolutionary, the
future of virtually all software

28
 6 Reasons for SaaS
1. No install worries about HW, OS
2. No worries about data loss (data is remote)
3. Easy for groups to interact with same data
4. If data is large or changes frequently,
simpler to keep 1 copy at central site
5. 1 copy of SW, single HW/OS environment
=> no compatibility hassles for developers
=> beta test new features on 1% of users
6. 1 copy => simplifies upgrades for
developers and 29
 SaaS Loves Rails
 Many frameworks/languages for SaaS
 Django/Python, Zend/PHP, Spring/Java
 We use Ruby on Rails (“Rails”)
 Rails popular programming framework that
uses Ruby – e.g., Twitter


30
 Why Take Time for Ruby/Rails?
 14 weeks to learn:
 SaaS, Agile, Pair Programming, Behavior
Driven Design, LoFi UI, Storyboards, User
Stories, Test Driven Development, Enhance
Legacy Code, Scrum, Velocity, JavaScript,
Design Patterns, UML, Deployment, Security
 You only work part-time on this class
 Only hope is highly productive language,
tools, framework: Rails currently the best
 See “Crossing the Software Education Chasm,”
A. Fox, D. Patterson, Comm. ACM
31
END
32
Which is WEAKEST Argument for a
Google App’s Popularity as SaaS?

☐ Don’t lose data: Gmail

☐ Cooperating group: Drive

this☐ Large/Changing Dataset: YouTube

☐ No field upgrade to get improved app:

Search

33
END
34
Service Oriented Architecture
(SOA)

(Engineering Software as a Service §1.4)

35
 Software Architecture
 Can you design software so that you can
recombine independent modules to offer
many different apps without a lot of
programming?

36
 Service Oriented Architecture
 SOA: SW architecture where all
components are designed to be services
 Apps composed of interoperable services
 Easy to tailor new version for subset of users
 Also easier to recover from mistake in design
 Contrast to “SW silo” without internal APIs

37
 CEO: Amazon shall use SOA!
1. “All teams will henceforth expose their data and
functionality through service interfaces.”
2. “Teams must communicate with each other
through these interfaces.”
3. “

38
 CEO: Amazon shall use SOA!
4. “It doesn't matter what [API protocol] technology
you use.”
5. “Service interfaces, without exception, must be
designed from the ground up to be
externalizable. That is to say, the team must plan
and design to be able to expose the interface to
developers in the outside world. No exceptions.”
6. “Anyone who doesn't do this will be fired.”
7. “Thank you; have a nice day!

39
 Bookstore: Silo

 Internal subsystems
can share data directly reviews users orders

 Review access user

profile Review User Profile Buying

 All subsystems inside

Subsystem Service Subsystem

single API
(“Bookstore”) Bookstore Service

(Figure 1.7, Engineering Software as a

Service by Armando Fox and David
Patterson, 1st edition, 2014.)
40
 Bookstore: SOA credit card
processing

 Subsystems user editor

users orders
reviews reviews
independent,
as if in separate
Review User Profile Buying
data centers Service Service Service

 Review
Service
accesses User Bookstore Service

Profile API
 Can recombine to
Favorite Books
make new service Service
users
(“Favorite Books”)
Social
(Figure 1.7, Engineering Software as a Network
Service by Armando Fox and David Service
Patterson, 1st edition, 2014.)
41
END
42
Which is NOT True About SOA?

SOA does not affect performance

this☐

Silo’ed systems are likely completely

☐ down on a failure, SOA must handle
partial failures
☐ SOA improves developer productivity
primarily through reuse
No SOA service can name or access
☐ another service’s data; it can only make
requests for data thru an external API
43
END
44
 Cloud Computing

(Engineering Software as a Service §1.6)

45
 What is Ideal HW for SaaS?
 Amazon, Google, Microsoft … developed
hardware systems to run SaaS
 What did they use: Mainframes?
Supercomputers?
 How can independent SW developers build
SaaS apps and compete without similar HW
investments to Amazon, Google, Microsoft?

46
 SaaS Infrastructure?
 SaaS’s 3 demands on infrastructure:
1. Communication
 Allow customers to interact with service
2. Scalability
 Fluctuations in demand during day + new
services to add users rapidly
3. Dependability


47
 Services on Clusters
 Clusters: Commodity computers connected
by commodity Ethernet switches
1. More scalable than conventional servers
2. Much cheaper than conventional servers
 20X for equivalent vs. largest servers
3. Dependability via extensive redundancy
4. Few operators for 1000s servers
 Careful selection of identical HW/SW
 Virtual Machine Monitors simplify operation
48
 Warehouse Scale Computers
 Clusters grew from 1000 servers to 100k
based on customer demand for SaaS apps
 Economies of scale pushed down cost of
largest data center by factors 3X to 8X
 Purchase, house, operate 100K v. 1K computers
 Traditional data centers utilized 10% - 20%
 Earn $ offering pay-as-you-go use at less
than customer’s costs for as many
computers as customer needs
49
 Utility Computing /
Public Cloud Computing
 Offers computing, storage, communication
at pennies/hour
 No premium to scale:
1000 computers @ 1 hour
= 1 computer @ 1000 hours
 Illusion of infinite scalability to cloud user
 As many computers as you can afford


50
 2013 AWS Instances & Prices
Instance Type $ Per $ Ratio Virtual Compute Memory Storage (GB)
Hour to small Cores Units (GB)
m1.small 0.06 1.0 1 1.0 1.7 1 x 160
m1.medium 0.12 2.0 1 2.0 3.8 1 x 410
m1.large 0.24 4.0 2 4.0 7.5 2 x 420
m1.xlarge 0.48 8.0 4 8.0 15.0 4 x 420
m3.xlarge 0.50 8.3 4 13.0 15.0 EBS
m3.2xlarge 1.00 16.7 8 26.0 30.0 EBS
c1.medium 0.15 2.5 2 5.0 1.7 1 x 350
c1.xlarge 0.58 9.7 8 20.0 7.0 4 x 420
cc2.8xlarge 2.40 40.0 32 88.0 60.5 4 x 840
m2.xlarge 0.41 6.8 2 6.5 17.1 1 x 420
m2.2xlarge 0.82 13.7 4 13.0 34.2 1 x 850
m2.4xlarge 1.64 27.3 8 26.0 68.4 2 x 840
cr1.8xlarge 3.50 58.3 32 88.0 244.0 2 x 120 SSD
hi1.4xlarge 3.10 51.7 16 35.0 60.5 2 x 1024 SSD 51
 Supercomputer for Hire
 Top 500 supercomputer competition in 2012
 532 8xlarge (@ $2.40/hour), 17000 cores =
240 TeraFLOPS
 72nd/500 supercomputer @ ~$1300 per hour
 Credit card => can use 1000s computers
 FarmVille on AWS
 Prior biggest online game 5M users
 What if startup had to build data center?
 4 days =1M; 2 months = 10M; 9 months = 75M
52
 IBM Watson for Hire?
 Jeopardy Champion IBM Watson
 Hardware: 90 IBM Power 750 servers
 3.5 GHz 32 cores/server
 360 @~$2.40/hour = ~$800/hour?
 Cost of top human attorney or accountant
 For what tasks could AI be as good as
highly trained person @ $800/hour?
 And rapidly falling cost

53
END
54
Which is NOT True About SaaS,
SOA, and Cloud Computing?
Clusters are collections of commodity
☐
servers connected by LAN switches
The Internet supplies the
☐
communication for SaaS apps
Cloud computing uses HW clusters + SW
☐
layer using redundancy for dependability
Private data centers could match cost of
☐ Warehouse Scale Computers if they just
used same type of HW and SW
55
END
56
Legacy SW vs. Beautiful SW

(Engineering Software as a Service §1.7)

57
 Programming Aesthetics
 Do I care what others think of my code?
 If it works, does it matter what code looks like?

58
 Legacy SW vs. Beautiful SW
 Legacy code: old SW that continues to
meet customers' needs, but difficult to
evolve due to design inelegance or
antiquated technology
 ___% SW maintenance costs adding new
functionality to legacy SW
 ___% for fixing bugs
 Contrasts with beautiful code: meets
customers' needs and easy to evolve

59
 Legacy SW vs. Beautiful SW
 Legacy code: old SW that continues to
meet customers' needs, but difficult to
evolve due to design inelegance or
antiquated technology
 60% SW maintenance costs adding new
functionality to legacy SW
 17% for fixing bugs
 Contrasts with beautiful code: meets
customers' needs and easy to evolve

60
 Legacy Code: Vital but Ignored
 Missing from traditional SWE courses and
textbooks
 #1 request from industry experts we asked:
What should be in new SWE course?
 Save work by reusing existing code (e.g., open
source)
 Will have legacy lectures and programming
assignments later in course


61
END
62
Which Type of SW is Considered an
Epic Failure?

☐ Beautiful code

☐ Legacy code

Unexpectedly short-lived code

☐this

☐ Both legacy code and unexpectedly

short lived code

63
END
64
 Quality Assurance &Testing

(Engineering Software as a Service §1.8)

65
 Software Quality
 What is software quality, and how to we
assure it? (QA)
 V&V: What is the difference (if any) between
Verification and Validation?

66
 Software Quality
 Product quality (in general): “fitness for use”
 Business value for customer and manufacturer
 Quality Assurance : processes/standards
=> high quality products & to improve quality
 Software quality:
1. Satisfies customers’ needs - easy to use, gets
correct answers, does not crash, …
2. Be easy for developer to debug and enhance
 Software QA: ensure quality and improve
processes in SW organization
67
 Assurance
 Verification: Did you build the thing right?
 Did you meet the specification?
 Validation: Did you build the right thing?
 Is this what the customer wants?
 Is the specification correct?
 Hardware focus generally Verification
 Software focus generally Validation
 Testing to Assure Software Quality

69
 Testing
 Exhaustive testing infeasible
 Divide and conquer: perform different tests
at different phases of SW development
 Upper level doesn’
System or acceptance test: integrated program meets
its specifications
Integration test: interfaces between units have
consistent assumptions, communicate correctly
Module or functional test: across individual units
Unit test: single method does what was expected 70
 More Testing
 Black box vs. White Box testing
 Testing based on specs vs. on implementation
 Test Coverage: % of code paths tested
 Regression testing: automatically rerun old
tests to check that changes don’t break
what used to work
 Continuous Integration (CI) testing:
continuous regression testing on each code
check-in vs. later testing phase
71
END
72
Which is NOT True About Testing?

☐ With better test coverage, you are

more likely to catch faults
While difficult to achieve, 100% test
☐this
coverage insures design reliability
Each higher level test delegates more
☐
detailed testing to lower levels
Unit testing works within a single class
☐
and module testing works across classes
73
END
74
 Productivity: Conciseness,
Synthesis, Reuse, and Tools

(Engineering Software as a Service §1.9)

75
 Productivity
 50 years of Moore’s Law => 2X /1.5 years
 HW designs get bigger
 Faster processors and bigger memories
 SW designs get bigger
 Had to improve SW productivity
 4 techniques
1. Clarity via conciseness
2. Synthesis
3. Reuse
4. Automation and Tools
76
1. Syntax: shorter and easier to read
assert_greater_than_or_equal_to(a,7)
vs. ________________

77
1. Syntax: shorter and easier to read
assert_greater_than_or_equal_to(a,7)
vs. a.should be ≥ 7
2. Raise the level of abstraction:
 HLL programming languages vs. assembler
 Automatic memory management (Java vs. C)
 Scripting languages: reflection,
metaprogramming

78
 Synthesis
 Software synthesis
 BitBlt: generate code to fit situation & remove
conditional test
 Research Stage: Programming by example

79
 Reuse
 Reuse old code vs. write new code
 Techniques in historical order:
1. Procedures and functions
2. Standardized libraries (reuse single task)
3. Object oriented programming: reuse and
manage collections of tasks
4. Design patterns: reuse a general strategy
even if implementation varies

80
 Automation and Tools
 Replace tedious manual tasks with
automation to save time, improve accuracy
 New tool can make lives better (e.g., make)
 Concerns with new tools: Dependability,
UI quality, picking which one from several
 Good software developer must repeatedly
learn how to use new tools: lifetime learning
 Lots of chances in this course:
Cucumber, RSpec, Pivotal Tracker, …

81
END
82
Which is TRUE About Productivity?

Copy and pasting code is another

☐
good way to get reuse
Metaprogramming helps productivity
☐true
via program synthesis
Of the 4 productivity reasons, the
true
primary one for HLL is reuse
A concise syntax is more likely to have
true
fewer bugs and be easier to maintain
83
END
84
 DRY
 “Every piece of knowledge must have a
single, unambiguous, authoritative
representation within a system.”
 Andy Hunt and Dave Thomas, 1999
 Don't Repeat Yourself (DRY)
 Don’t want to find many places have to apply
same repair
 Refactor code so that it has a single place to
do things
85
END
86
And in Conclusion: §§1.1, 1-4-1.9
 Class: SW eng. Principles via Cloud app by team
for customer + enhancing legacy app
 SaaS less hassle for developers and users
 Service Oriented Architecture makes it easy to
reuse current code to create new apps
 Scale led to savings/CPU => reduced cost of
Cloud Computing => Utility Computing
 Testing to assure software quality, which means
good for customer and developer
 Developer Productivity: Conciseness, Synthesis,
Reuse, and Tools
87