Software Development Methodologies Lecturer: Raman Ramsin Lecture - - PowerPoint PPT Presentation

software development methodologies
SMART_READER_LITE
LIVE PREVIEW

Software Development Methodologies Lecturer: Raman Ramsin Lecture - - PowerPoint PPT Presentation

Software Development Methodologies Lecturer: Raman Ramsin Lecture 10 Agile Methodologies: XP Sharif University of Technology Department of Computer Engineering 1 Software Development Methodologies Lecture 10 eXtreme Programming (XP)


slide-1
SLIDE 1

Software Development Methodologies

Lecturer: Raman Ramsin Lecture 10 Agile Methodologies: XP

Department of Computer Engineering

1

Sharif University of Technology

slide-2
SLIDE 2

Software Development Methodologies – Lecture 10

eXtreme Programming (XP) eXtreme Programming (XP)

Developed by Beck in 1996 Developed by Beck in 1996. The first authentic XP book appeared in 1999 with a The first authentic XP book appeared in 1999, with a

revised and refined version appearing in 2004.

Although some of the methodologies that are nowadays

dubbed as agile are older than XP, it was the advent of XP that sparked the agile movement XP that sparked the agile movement.

XP considers itself a software engineering discipline XP considers itself a software engineering discipline

rather than a methodology, yet it does incorporate a process.

Department of Computer Engineering

2

Sharif University of Technology

slide-3
SLIDE 3

Software Development Methodologies – Lecture 10

XP: Process

  • 1. Exploration:
  • 1. developing an initial list of high-level requirements

2 d t i i th ll d i f th t th h t t i

  • 2. determining the overall design of the system through prototyping
  • 2. Planning (Release Planning):
  • 1. estimating the time needed for the implementation of each requirement

2 h

  • 2. prioritizing the requirements
  • 3. determining a schedule and a minimal, select set of requirements for the first release of

the system

3 Iterations to First Release: iterative development of the first release of the system

  • 3. Iterations to First Release: iterative development of the first release of the system

using the specific rules and practices prescribed by XP; iterations are typically between 1 to 3 weeks in duration. 4 Productionizing: system wide verification and validation of the first release and its

  • 4. Productionizing: system-wide verification and validation of the first release, and its

deployment into the user production environment

  • 5. Maintenance: implementing the remaining requirements (including any resulting from

post deployment maintenance needs) into the running system by iterating phases 2 post-deployment maintenance needs) into the running system by iterating phases 2, 3 and 4.

  • 6. Death: closing the project and conducting post-mortem review and documentation

Department of Computer Engineering

3

Sharif University of Technology

slide-4
SLIDE 4

Software Development Methodologies – Lecture 10

XP: Process

Department of Computer Engineering

4

Sharif University of Technology

[Abrahamsson et al. 2002]

slide-5
SLIDE 5

Software Development Methodologies – Lecture 10

XP: Process – Development Engine p g

The activities performed in the second, third and fourth phases

  • f the process constitute the development “engine” of the XP

methodology. methodology.

Each execution (run) of these phases produces a new release. A first release of the system is initially produced and deployed,

which is then incrementally improved and complemented which is then incrementally improved and complemented during the maintenance phase through further iterations (runs)

  • f the development engine.

Department of Computer Engineering

5

Sharif University of Technology

slide-6
SLIDE 6

Software Development Methodologies – Lecture 10

XP: Process

Department of Computer Engineering

6

Sharif University of Technology

[Ambler 2002]

slide-7
SLIDE 7

Software Development Methodologies – Lecture 10

XP Process: Exploration p

  • 1. Formation of the development team: the team typically consists of

A coach acting as monitor and facilitator, a number of programmers, and a business

representative (customer) participating in project activities and supplying the team representative (customer) participating in project activities and supplying the team with information and feedback

[Optional] A number of analysts to help elicit the requirements, a number of testers

helping define acceptance tests, and a resource manager

  • 2. Development of the initial set of User Stories: A User Story

defines a system feature as seen from the customer’s point of view. is written by the customer in his own terminology on index cards.

y gy

Is nothing but a short description (around three sentences) of a certain chunk of

functionality needed to be delivered by the system.

  • 3. Creation of the system Metaphor:

y p A prototype (called Spike or Spike Solution) is developed, exploring potential

architectures for the system.

The prototype helps the team define the system Metaphor, which is typically a very

simple high-level description of how the system works simple, high level description of how the system works.

It usually takes the form of a description-by-analogy in order to be easily

understandable to all the team members.

Though informal, the metaphor gives an extremely useful idea of the overall

hit t f th t ith t tti t t i t

Department of Computer Engineering

7

Sharif University of Technology architecture of the system without setting too many constraints.

slide-8
SLIDE 8

Software Development Methodologies – Lecture 10

XP Process: Development Engine – Top-Level Activities p g p

Department of Computer Engineering

8

Sharif University of Technology

[Wells 2003]

slide-9
SLIDE 9

Software Development Methodologies – Lecture 10

XP Process: Planning (Release Planning) g ( g)

  • 1. Estimation of development time:
  • 1. Developers estimate the time needed to develop each of the user stories as

p p conforming to the system metaphor, and write the estimates down on the user-story index cards.

  • 2. User stories that need more than 3 weeks to develop are broken down into

ll d t i t ki l th 1 k d smaller ones, and user stories taking less than 1 week are merged.

  • 3. In cases where estimates are not reliable enough, spike solutions are

developed in order to improve the estimates.

  • 2. Prioritization of user stories: the customer prioritizes the user stories according

to their business value. 3 Pl i th fi t l

  • 3. Planning the first release:
  • 1. the team selects a minimal, most valuable set of user stories for

implementation in the first release, and agrees on the release date. 2 Th t l d id th it ti d ti (b t 1 t 3 k )

  • 2. The team also decides on the iteration duration (between 1 to 3 weeks),

which once determined, will be the same for all iterations.

Department of Computer Engineering

9

Sharif University of Technology

slide-10
SLIDE 10

Software Development Methodologies – Lecture 10

XP Process: Iterations to First Release

  • 1. Iteration planning: at the start of each iteration, a planning meeting is held,

performing:

  • 1. Selection of user stories to implement, and failed acceptance tests to rectify

p , p y

  • 2. Identification of programming tasks
  • 3. Task sign-up and estimation
  • 2. Development: the development activity in each iteration is itself an iterative

process with daily cycles consisting of the following activities:

  • 1. Holding daily stand up meetings

2 l d d d C ll C d O h

  • 2. Analysis, design, coding, testing and integration in a Collective-Code-Ownership

environment

Department of Computer Engineering

10

Sharif University of Technology

slide-11
SLIDE 11

Software Development Methodologies – Lecture 10

XP Process: Development Engine – Iteration Activities p g

Department of Computer Engineering

11

Sharif University of Technology

[Wells 2003]

slide-12
SLIDE 12

Software Development Methodologies – Lecture 10

XP Process: Iterations to First Release – Iteration Planning

  • 1. Selection of user stories to implement and failed acceptance tests to rectify:

based on the release plan, the customer selects user stories (according to their based on the release plan, the customer selects user stories (according to their

business value) for development in the coming iteration.

Failed acceptance tests encountered during previous iterations are also

considered for inclusion in the list of jobs to be attended to.

S

i l tt ti i i t th i i d d i i it ti

Special attention is given to the experience gained during previous iterations as

to the team's development speed (Project Velocity), to ensure that the selected jobs can be completed in the iteration. 2 Identification of programming tasks:

  • 2. Identification of programming tasks:

the developers on the team break down the selected user stories and

debugging jobs into programming tasks.

Tasks are then written down on the user-story index cards. Tasks are then written down on the user story index cards.

  • 3. Task sign-up and estimation:

programmers sign-up to do the tasks. Each developer estimates the time needed for completion of each of the tasks Each developer estimates the time needed for completion of each of the tasks

undertaken, making sure that all of them can be developed in the time

  • available. Each task should take between 1 to 3 days to complete.

Department of Computer Engineering

12

Sharif University of Technology

slide-13
SLIDE 13

Software Development Methodologies – Lecture 10

XP Process: Iterations to First Release - Development p

  • 1. Holding daily stand up meetings:

A short stand up meeting is held every morning in order to communicate A short stand up meeting is held every morning in order to communicate

problems and solutions, and help the team keep on track.

  • 2. Analysis, design, coding, testing and integration in a Collective-Code-

Ownership environment: Ownership environment:

Collective Code Ownership means that all the code developed is put in a shared

code repository, and any developer can change his or others’ code in order to add functionality, fix bugs, or refactor. y, g ,

test-driven development to make collective code ownership possible:

the developers have to create unit tests for their code as they develop it. All the code in the code repository includes unit tests, forming a suite of tests that is

automatically applied by test tools whenever code is added or changed automatically applied by test tools whenever code is added or changed.

The test suite safeguards the repository from malignant change: for code to be

allowed integration into the repository, it must pass the entire test suite.

Black-box acceptance tests based on the user stories are defined by the customer

and developed by the team during the iteration and are frequently applied (by and developed by the team during the iteration, and are frequently applied (by automated tools) to the code.

Department of Computer Engineering

13

Sharif University of Technology

slide-14
SLIDE 14

Software Development Methodologies – Lecture 10

XP Process: Development Engine – Iteration Development Activities

Department of Computer Engineering

14

Sharif University of Technology

[Wells 2003]

slide-15
SLIDE 15

Software Development Methodologies – Lecture 10

XP Process: Development Engine – Iteration Development Activities Activities in the Collective-Code-Ownership Environment

  • Programmers work in pairs, each pair on one machine (a practice called

Pair Programming).

  • Programmers use CRC cards in order to come up with the simplest design

g p p g possible for the programming task in hand.

  • Refactoring is constantly done in order to simplify the code and eliminate

redundancy.

  • A common coding standard is enforced in order to promote code legibility,

which in turn enhances communication among developers.

  • Developers are moved around so that they acquire knowledge about all

t f th t thi ill d th t f h d t th t parts of the system; this will reduce the cost of changes made to the team structure and will help relieve overloading and coding bottlenecks.

  • Builds are frequent in XP, and continuous integration is encouraged.

D l t k t t i bl ith f t h k

  • Developers are to work at a sustainable pace, with forty hours a week as

the norm; nobody is allowed to work overtime for two weeks in a row.

Department of Computer Engineering

15

Sharif University of Technology

slide-16
SLIDE 16

Software Development Methodologies – Lecture 10

XP Process: Development Engine – Iteration Development Activities Activities in the Collective-Code-Ownership Environment

Department of Computer Engineering

16

Sharif University of Technology

[Wells 2003]

slide-17
SLIDE 17

Software Development Methodologies – Lecture 10

XP Process: Productionizing

  • 1. System-wide verification and validation: The release is tested to make sure of

the user’s approval and the system’s readiness for deployment.

Acceptance tests mostly developed during the iterations-to-first-release phase Acceptance tests, mostly developed during the iterations to first release phase,

are used here as regression tests.

Defects found are resolved through iterations of the main development cycle.

  • 2. Deployment into the production environment: the release is introduced into

the user environment.

Involves the usual integration conversion tuning training and documentation Involves the usual integration, conversion, tuning, training, and documentation

activities typical of deployment efforts.

Any tuning and stabilization action on the release itself is regarded as a

development activity and is conducted through short iterations (typically weekly)

  • f the development cycle
  • f the development cycle.

Department of Computer Engineering

17

Sharif University of Technology

slide-18
SLIDE 18

Software Development Methodologies – Lecture 10

XP Process: Maintenance

  • The maintenance phase is when the remaining user stories are implemented into

the system and the system is maintained as such the system and the system is maintained as such.

  • Encompasses the same activities as those in the phases of the development engine;

i.e. Planning, Iterations to Release, and Productionizing. g, , g

  • Still dependent on the evolving set of user stories and the system metaphor.
  • The small releases produced during maintenance are integrated into an already

running and operational system. Req i ements a ising as a es lt of maintenance a e t eated as o dina

  • Requirements arising as a result of maintenance are treated as ordinary

requirements (expressed as user stories) and implemented through the same iterative development process.

  • The maintenance phase continues until either there are no more user-stories to

develop and none are anticipated in the future, or the system in no way lends itself to necessary evolution any more.

Department of Computer Engineering

18

Sharif University of Technology

slide-19
SLIDE 19

Software Development Methodologies – Lecture 10

XP Process: Death

The project is declared dead when evolution is either unnecessary or

impossible.

The main activities performed in this phase of the XP process are:

  • 1. Declaring the project as closed: this involves wrapping up the usual legal,

fi i l d i l l d financial and social loose ends.

  • 2. Post-mortem documentation and review: preparing a short document (no longer

than ten pages) providing a brief tour of the system, and writing a review i i h l l d f h j report summarizing the lessons learned from the project.

Department of Computer Engineering

19

Sharif University of Technology

slide-20
SLIDE 20

Software Development Methodologies – Lecture 10

XP: Strengths and Weaknesses

  • Strengths
  • Iterative-incremental process
  • Based on system functionality captured in User Stories
  • Based on system functionality captured in User Stories
  • The process is tuned according to feedback during its

execution execution

  • Traceability to requirements through the use of user stories

throughout the process as the basis for tasks and tests

  • Based on system architecture (Metaphor) identified through

prototyping A ti i l t

  • Active user involvement
  • Test-based development

Department of Computer Engineering

20

Sharif University of Technology

slide-21
SLIDE 21

Software Development Methodologies – Lecture 10

XP: Strengths and Weaknesses

  • Strengths (Contd.)
  • Stringent standards enforced on coding
  • Early and frequent releases
  • Early and frequent releases
  • Requirements are allowed to evolve over time
  • Iterative development engine governed by careful planning and

p g g y p g reviewing

  • Explicit coverage of maintenance and project retirement

(Death) phases; maintenance in fact comprises the bulk of the (Death) phases; maintenance in fact comprises the bulk of the development effort

  • Continuous validation
  • Continuous integration
  • Refactoring exercised in order to acquire the simplest code

possible

Department of Computer Engineering

21

Sharif University of Technology

possible

slide-22
SLIDE 22

Software Development Methodologies – Lecture 10

XP: Strengths and Weaknesses

  • Weaknesses
  • Process is rather vague: the process commonly introduced as

the “XP Process” is just a typical example the XP Process is just a typical example.

  • More intended as a set of principles and practices rather than

a methodology gy

  • Limited evidence of scalability
  • Seamlessness is not addressed: development is more or less a

jump from user stories to code, and the little design that is done (if at all) does not have to conform to any standard.

Department of Computer Engineering

22

Sharif University of Technology

slide-23
SLIDE 23

Software Development Methodologies – Lecture 10

XP: Strengths and Weaknesses

  • Weaknesses (Contd.)
  • Requires the use of automated tools and enforcement of

discipline for “Collective-Code-Ownership” to be practicable. p p p

  • No clear-cut design effort
  • Model-phobic
  • Except for CRC cards models are not prescribed leaving it to
  • Except for CRC cards, models are not prescribed, leaving it to

the individual developer to decide what model is useful to him.

  • Lack of formalism

Department of Computer Engineering

23

Sharif University of Technology

slide-24
SLIDE 24

Software Development Methodologies – Lecture 10

References

  • Abrahamsson P

Salo O Ronkainen J Warsta J Agile Software

  • Abrahamsson, P., Salo, O., Ronkainen, J., Warsta, J., Agile Software

Development Methods: Review and Analysis. VTT Publications, 2002.

  • Wells D

Extreme programming: A gentle introduction Published on the

  • Wells, D., Extreme programming: A gentle introduction. Published on the

Web at: http://www.extremeprogramming.org, 2003, visited in April 2006.

  • Beck K

and Andres C Extreme Programming Explained: Embrace

  • Beck, K., and Andres, C., Extreme Programming Explained: Embrace

Change, 2nd ed. Addison-Wesley, 2004.

  • Ambler S W

AM Throughout the XP Lifecycle Published on the Web at:

  • Ambler, S. W., AM Throughout the XP Lifecycle, Published on the Web at:

http://www.agilemodeling.com/essays/agileModelingXPLifecycle.htm, 2002, visited in

April 2006.

Department of Computer Engineering

24

Sharif University of Technology