The Agile V ModelOxymoron or Best Practice?
I t d tiIntroductionThe V model: a dead relic of deluded The V model: a dead relic of deluded waterfallers?Advocates of the V model: anachronistic di ? dinosaurs? Actually, the V model was ahead of its time in many waysmany waysLet’s look at the way the debate is framed, then smash a few mythsI’ll review a number of best practices of agile development that are actually V model best practices
The Agile V Model Page 2
p act ces
www.rbcs-us.com Copyright (c) RBCS 2014
Th S ti l V M d l?The Sequential V Model?
…but is the V model always necessarily sequential? Can we
t d it id
The ISTQB syllabi and most test experts (pro and con)
id th V d l
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 3
extend its core ideas into agile?
consider the V model sequential…
A il L k LikAgile Looks Like…
Yeah, the model looks totally different than the V d l ’ d f d l dmodel. However, I’ve identified nine test-related
agile concepts (so far) that are clearly rooted in, derived from, or related to V model best practices.
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 4
, p
A il B t P ti f th V M d lAgile Best Practices from the V ModelAgile best practices rooted in V model ideas:Agile best practices rooted in V model ideas:
Testing and quality throughout iterationWhole team approachKill b h Kill bugs at the sourceAcceptance criteriaAcceptance test driven development and behavior p pdriven developmentDefinition of doneTest quadrants qPlanning pokerBurndown charts and velocity
D ’t b li ? I’ll h Don’t believe me? I’ll show you…
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 5
T ti d Q lit Th h tTesting and Quality ThroughoutAgile emphasizes testing and quality Agile emphasizes testing and quality throughout each iteration, not just at the end
Working software is the primary measure of progressEmphasize technical excellence and good design
Consider the V model principle of early Consider the V model principle of early testing and quality assurance
User story grooming at start of each iterationTest driven development
Both are examples of activities related to this principleprinciple
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 6
Wh l T App hWhole Team ApproachIt’s an agile best practice g pto use the whole team approachThis means developers, testers and business testers, and business stakeholders collaborate throughout the process
This collaboration is essential to testing and QA throughoutIt also promotes efficiency, teamwork, and high , gmorale
The V model calls for same collaboration, for same reasonssame reasons
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 7
Kill B t th SKill Bugs at the SourceAn agile best practice is to fix bugs in the iteration in An agile best practice is to fix bugs in the iteration in which they are found
“Fix bugs first in next iteration” is seen as a weaker strategyHardening or stabilization sprints seen as waterfall by Hardening or stabilization sprints seen as waterfall by purists
This concept is similar to phase containment, a V model best practicemodel best practice
Phase containment seeks to increase quality and reduce costs by removing defects at point of introductionAgile seeks to increase quality and reduce costs by removing Agile seeks to increase quality and reduce costs by removing defects in the iteration of introduction
Lifecycle aside, the song remains the same: it’s cheaper and better to find ‘em and fix ‘em soonercheaper and better to find em and fix em sooner
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 8
A pt C it iAcceptance CriteriaUser stories (the usual agile version of User stories (the usual agile version of requirements) are supposed to include acceptance criteria Acceptance criteria tell developers, testers, and business stakeholders:
What will be tested (the test conditions)What will be tested (the test conditions)How to recognize successful implementation (the test oracle)
Identifying test conditions and test oracles before coding (i.e., test analysis before code) is a long standing V model best practiceis a long-standing V model best practice
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 9
ATDD d BDDATDD and BDDAn agile best practice is the elaboration of user An agile best practice is the elaboration of user stories and acceptance criteria into tests
Acceptance Test Driven Development (ATDD)/Specification by ExampleBehavior Driven Development (BDD)
ATDD/BDD provide specific test cases readable ATDD/BDD provide specific test cases, readable by business and technical stakeholders, that show how a feature should workCreating test cases before coding (i.e., test design before code) is another long-standing V model b ibest practice
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 10
D fi iti f DDefinition of DoneIn Scrum (a dominant agile test management In Scrum (a dominant agile test management process), participants create a “definition of done” for:
Process activitiesProcess activitiesWork products
Promotes quality and efficiencyVery similar to the form and motivation behind a Very similar to the form and motivation behind a V model best practice, the definition of exit criteriaH i th V d l it ( d t ) it i However, in the V model, exit (and entry) criteria are often used to enforce sequentiality, a no-no in agile
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 11
T t Q d tTest QuadrantsAn agile best practice Business orientedAn agile best practice is to use the test quadrants to identify the right tests at the y)
Evalu
Mixed ManualSystem level:
functionality, test System/UAT level: gright time in each iterationThis maps directly to ea
m (v
erify
uate the pr
y,acc. criteria, canbe included in
regression tests
y /reactive tests,
scenarios, usability
This maps directly to the V model concept of test levels with varying focus and po
rt th
e te roduct (va
Unit level: unit/ component
tests in continuous
System/OAT level:
reliability, perf., d t i it a y g ocus a d
owners……but again without sequentiality
Supp
alidate)
Automated Automated/Tools
integration process
data mig., security, maintainability,
comp., interop, etc.
sequentiality
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 12
Technology oriented/
Pl i P kPlanning PokerPlanning poker is a Scrum estimation practice Planning poker is a Scrum estimation practice using story points, t-shirt sizes, or person-hours
Each member of the team estimates task size privately in a group meetingAll reveal their estimates at once (avoids All reveal their estimates at once (avoids anchoring)Discussion is used to build consensus if needed
Thi i thi th th D l hi l This is nothing more than the Delphic oracle estimation technique, used for years in V model projectsmodel projects
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 13
B d Ch t d V l itBurndown Charts and VelocityAfter estimation, ,burndown charts are used to track work completed
This gives the team’s true velocity (capability for completion)Velocity is used to Velocity is used to estimate future iterations
This is a form of metrics-based
lestimation, a long-standing V model best practice
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 14
What to Lose and to Keep in an Agile VWhat to Lose and to Keep in an Agile V
Lose KeepLoseSequentiality (with its enabler, limiting
KeepIndependent testers(Flexible) up-front d
, gchange)Delivery only at the
( ) pdesignTasks outside iterations (e.g., automation, perf
endAny adversarialism
IV&V ( )
( g , , ptest, data, env, etc.)Other lifecycles where neededIV&V (process cop)
Testers approving releases (quality cop)
neededLifecycle tailoringDefect tracking
The Agile V Modelwww.rbcs-us.com
Copyright (c) RBCS 2014 Page 15
C l iConclusionsThe V model is dead…long live the V model!gIn fact, sequential lifecycles (as well as other lifecycles) still have a placeMany agile best practices arose from V model ideasMany agile best practices arose from V model ideasToo bad that too few teams ever really followed these V model best practices (until now, under the guise of agile)guise of agile)Escape from purist dogmas to embrace good ideas where you find them, in any lifecycleD li ( ith / ) i ld f l t l Dualism (either/or) is seldom a useful mental model; usually it’s an obstacle Instead: tailor, adapt, blend, and extend
The Agile V Model Page 16www.rbcs-us.com
Copyright (c) RBCS 2014
T C t t RBCSFor 20 years, RBCS has delivered consulting, outsourcing and training services to clients helping them with software and hardware testing Employing the
To Contact RBCSclients, helping them with software and hardware testing. Employing the industry’s most experienced and recognized consultants, RBCS advises its clients, trains their employees, conducts product testing, builds and improves testing groups, and hires testing staff for hundreds of clients worldwide. Ranging from g p g g gFortune 20 companies to start-ups, RBCS clients save time and money through improved product development, decreased tech support calls, improved corporate reputation and more. To learn more about RBCS, visit www.rbcs-us.com.Add RBCS IAddress: RBCS, Inc.
31520 Beck RoadBulverde, TX 78163-3911USA
Phone: +1 (830) 438-4830E-mail: [email protected]: www.rbcs-us.comT itt @RBCS
The Agile V Model Page 17
Twitter: @RBCSFacebook: RBCS-Inc
www.rbcs-us.com Copyright (c) RBCS 2014