User Stories

Testing 2, October 21
 Admin
• Grades posted for HW 1 – stop by 238 if
you want an explanation
• Don’t embed magic numbers in names
• Proj 1 due Saturday 8AM sharp
• HW #2 due Thursday 10/28 3:30PM
• Magic tournament this weekend
(http://unityentertainment.net)
 Proj 1 Proposed Weighting
• 20% functionality (2 points/story)
• 40% tests (green bar on submittal, test
design, test coverage)
• 20% refactoring
• 20% documentation (iterations,
retrospective, test case list)
 User Stories
• 3 aspects
– Written description of story used for
planning/reminding
– Conversations about the story that serve to fill
out details
– Tests that convey and document details & can
be used to determine when a story is
complete
• Represent functionality that will be valued
by users
 Examples of specifying value to
users
• Good:
– “A user can search for jobs”
– “A company can post new jobs”
• Bad:
– “The software will be written in C++”
– “The program will connect to the database
through a connection pool”
 Sizing stories
• Too broad = impossible to test/code
• Too narrow = more time spent specifying than
implementing
• Aim for test & code cycle of about 4 hours to 2
weeks by one or 2 programmers per story
• Split long stories (“epics”) into smaller pieces
• Rather than specify small details, get those in
conversations with customer & annotate story
• Big stories can serve as placeholders for areas
of the system that still need to be discussed
 Writing Stories
• Customer writes stories
– Written in language of business to allow prioritization
– Customer is primary product visionary
• Good stories are
– Independent
– Negotiable
– Valuable to users or customers
– Estimatable
– Small
– Testable (INVEST)
 Indepent
• Stories that depend on other stories are
difficult to prioritize and estimate
• Example:
– A company can pay for a job posting with a
Visa card
– A company can pay for a job posting with a
Mastercard
– A company can pay for a job posting with an
American Express card
 Negotiable
• Story cards serve as reminders not
contracts
• Details need to be fleshed out in
conversation
• Story cards should have a phrase or
sentence to serve as reminder to have
conversation & notes about conversation
 Valuable
• Both to people using the software and paying for
the software
• Avoid stories valued only by developers (make
the benefits to customers/users apparent for
these stories)
• Example
– “All connections to the database are through a
connection pool” could be rewritten as “Up to 50
users should be able to use the application with a 5-
user database license”
 Estimable
• 3 common reasons why story might not be
– Developers lack domain knowledge
• Get details from customer
– Developers lack technical knowledge
• Perform spike to explore technology
– Story is too big
• Split the story into smaller ones
 Small
• Easy to use in planning
• Split compound & complex stories
• Combine too small stories
 Splitting Stories
• Compound
– Conversations may reveal multiple stories
– Split along Create/Update/Delete
– Split along data boundaries
• Complex
– Split into investigative and develop the new feature
stories (define timebox for investigation)
• Make sure each split-off portion is a good story
(INVEST)
 Testable
• Can’t tell if story is done without tests
• Aim for most tests to be automatable
Story Responsibilities (Developers)
• Help the customers write stories that
– Are promises to converse rather than detailed specs
– Have value to the users or the customer
– Are independent
– Are testable
– Are appropriately sized
• Describing the need for technology/infrastructure
in terms of value to users or customers
• Have the conversations with the customers
Story Responsibilities (Customers)
• Writing stories that
– Are promises to converse rather than detailed
specs
– Have value to users or to yourself
– Are independent
– Are testable
– Are appropriately sized
• Have the conversations with the
developers
 Users
• Can be more than one type of user for
system
• Different user types may have different
roles and different stories
• Disfavored users (e.g., hackers) may be
included as well as favored users
– Obviously, the value to users piece gets
flipped for disfavored users
 Role Modeling
• Brainstorm an initial set of user roles
• Organize the initial set
• Consolidate roles
• Refine the roles
Roles for Magic DB?
 Attributes worth considering when
refining roles
• Frequency with which user will use
software
• User’s level of expertise with domain
• User’s general level of proficiency with
computers and software
• User’s level of proficiency with this
software
• User’s general goal for using software
 Additional user modeling
• Identify personas
– Should be described sufficiently so everyone
on team feels like they know this “person”
– Choose personas that truly represent user
population
• Extreme characters
– Make a PDA for
• Drug dealer
• The Pope
• 20-year old woman juggling multiple boyfriends
 User Modeling Responsibilities
(Developers)
• Participate in process of identifying user
roles and personas
• Understanding each user role & how they
differ
• Thinking about how different user roles
might prefer their software to behave
• Identifying & describing user roles doesn’t
go beyond role of tool in the process
 User Modeling Responsibilities
(Customers)
• Looking broadly across space of possible users &
identifying appropriate roles
• Participating in process of identifying roles and personas
• Ensuring software does not focus inappropriately on a
subset of users
• Ensuring that each story can be associated with at least
one user role or persona
• Thinking about how each user role may prefer their
software to behave
• Identifying & describing user roles doesn’t go beyond
role of tool in process
 Getting stories
• Trawl for stories (rather than elicit)
– Different size nets
– Fish grow and die
– Won’t capture all fish by trawling
– Skill plays a factor
• Techniques
– User interviews
– Questionnaires
– Observation
– Story-writing workshops
Story Gathering Responsibilities
• Understand & use multiple techniques
• Knowing how to make use of open-ended and
context-free questions
• Customer-specific
– Write as many user stories as early as possible
– Understanding options in communicating with users
– Scheduling & running story-writing workshop
– Ensuring all user roles represented when trawling
 Acceptance Tests
• Use tests to track details
• Write the tests before coding
• Acceptance tests are ideally specified by
the customer
• Don’t replace unit tests
 Acceptance Test Responsibilities
(Developers)
• Automating the execution of acceptance
tests
• Thinking about additional acceptance tests
when starting development
• Unit testing
 Acceptance Test Responsibilities
(Customers)
• Writing acceptance tests
• Executing acceptance tests
 Guidelines for Good Stories
• Start with goal stories
• Write closed stories (stories that have a
definite end point)
– “A recruiter can review resumes from
applicants to one of her ads” instead of “A
recruiter can manage the ads she has placed”
• Put constraints on the system on cards &
implement automated tests for them
 Guidelines II
• Size your story appropriately for the time frame it
may be implemented in
• Keep the UI out as long as possible
• Don’t rely solely on stories if somethings are
better expressed in other ways
• Include user roles in stories rather than saying
“user”
• Write for a single user (“A Job Seeker” not “Job
Seekers”
• Use Active Voice
 Son of Guidelines
• Don’t number the story cards
• Don’t forget the purpose of using story
cards
 Using Stories
• Assign points to stories based on difficulty
of coding
• For each release, customer prioritizes
stories
• Developers determine velocity (# of points
per release cycle) for previous cycle and
plan to implement the highest priority
stories up to that number of points for the
release
 User Stories are Not
• Requirements documents
• Use Cases
• Scenarios
 Why do user stories?
• Emphasize verbal communication
• Comprehensible by everyone
• Right size for planning
• Work for iterative development
• Encourage deferring detail
• Support opportunistic development
• Encourage participatory design
• Build up tacit knowledge
 Why not user stories?
• May be difficult to understand relations
between stories
• Not suitable for requirements traceability
(if required by process)
• May not scale well to large teams
 Story Smells
• Stories are too small
• Interdependent stories
• Goldplating
• Too Many details
• Including UI detail too soon
• Thinking too far ahead
• Splitting too many stories
• Customer has trouble prioritizing
• Customer won’t write and prioritize the stories