What are we doing today?
Agile Game Development Tutorial • Defining “agile” • Describe two agile methodologies • Scrum (this morning) • XP (this afternoon) Introduction • Agile for non-programmers • Scrum evolved for programmers • XP coevolved to support engineering in an agile environment • What will we need for other disciplines? • Discuss how agile alters roles for Clinton Keith • Producers, artists and designers CTO, • Engine/Middleware providers (Epic) • How agile planning works High Moon Studios
What are we doing today? Assumptions
• It’s just methodology • 30 minute presentations • It’s about identifying and removing the • 10 minutes of Q&A impediments for talented teams making great games • All levels of agile experience • We don’t compete based on process • Side-discussions encouraged • Poor process is hurting our industry • But well off to the side, so we don’t interrupt • It’s not easy. This will take a lot of effort and speakers sharing to get the full benefit
Schedule
Topic Start End What it means to be agile Introduction 10:00 AM 10:15 AM What is Agile? 10:15 AM 10:45 AM Break 11:00 AM 11:15 AM Mike Cohn Scrum 11:15 AM 11:45 AM Agile Game Design 11:55 AM 12:25 PM Lunch 12:35 PM 2:00 PM Agile Art Creation 2:00 PM 2:30 PM Agile Production 2:40 PM 3:10 PM Break 3:30 PM 3:45 PM Agile Planning 3:45 PM 4:15 PM XP/TDD 4:25 PM 4:55 PM Scrum & Middleware 5:05 PM 5:35 PM Closing comments 5:45 PM 6:00 PM
Handouts available
1 Is game development a A defined process defined process?
• Is every task completely understood? A defined • Are we even getting closer? process • Given the exact same inputs (including people) • Will we get the same results every time? • Can we even have the exact same inputs?
• Every task must be completely understood. • When given a well-defined set of inputs, the same outputs are generated every time.
Project noise level Empirical process control Project noise level Far from Agreement • Useful when Anarchy • Process cannot be sufficiently described to Complex ensure repeatability Complicated • There is so much complexity or noise that the process leads to different outcomes Source: Strategic Management and Organizational Dynamics by Ralph • Expects the unexpected
Requirements Stacey in Agile Software Development with Scrum by Ken Schwaber and Mike • Exercises control through frequent Close to Simple Beedle. Agreement inspection and adaptation Technology Close to Far from Far Certainty Certainty
Empirical model Three key words
Controls
Inputs Output Inspect and Adapt •Requirements • Potentially •Technology Process shippable •Team product increment
2 Agile is about shifting the The Agile Manifesto balance Individuals and Process and over Anticipation Interactions Tools
Working Comprehensive over software documentation
Contract Customer over Adaptation collaboration negotiation
Responding to Agile over Following a plan change
What is Scrum?
People and Scrum for game communication development Working game
Customer collaboration Clinton Keith CTO, Responding to High Moon Studios change
What is Scrum? How is it different? Product The Scrum Cast Owner Iterative Scrum Results oriented Not plan based Master Anyone. Not a Director lead role Collaborative Artist Artist The Animator Team Self disciplined Not management Marketing Commitment based QA oriented Programmer
About making things visible so you can apply common sense Designer Publisher Producer Customers
3 Product Backlog Sprints (Iteration)
Change Sprint Sprint Backlog (Fixed broken out by Duration) team Sprint Goals
Ga me
Jump Product Backlog as prioritized by Product Owner Jump New Game Crouch Crouch (Vertical Slice Added) Defined as User Stories with conditions of satisfaction Run Run Estimated with relative User Story Points that help track progress
Review and Planning Daily Scrum
Daily Scrum
Jump Ga me Jump Ga me
CrouchFly Fly
Run Review and Planning Crouch
Run
Jump User Story The War Room Task Mon. Tues. Wed. Thurs. Jump Input Control 8 3 0 0 Jump Tuning 16 16 10 4 Jump Animation 8 16 16 4 New task 8 8 0
50
40
30
20
10
0 Mon Tues Weds Thurs User Stories Tasks Completed Tasks (Sprint Backlog) Burndown Chart
4 Sprint Backlog Burndown Chart Scaling Scrum - The Scrum of Scrums
800 Functional Drag works against velocity Leadership 700 Support Slope = Velocity services 600 500 Hours 400 Backlogged 300 Story 200 100 0 Days Teams Must hit zero hours by end
Crunch Starting agile – some advice
Overtime Velocity Normal Velocity
Week 1 Week 2 Week 3 Week 4 Week 5
Agile Game Design
• Rory McGuire • Senior Game Designer Questions? • High Moon Studios
5 Design Department What are we talking about? At High Moon
• Agile Game Design At High Moon Studios • Define specific gameplay vision and participate in setting • Structure project goals. • Benefits • Own the creation of levels and mechanics. • Contribute ideas and enhancements to the game backlog. • Challenges • Are advocates for the player, helping to assess and focus feedback from testers and publishing partners.
Designers at High Moon What is a Customer?
• Focus on agile implementation of the • Projects are broken up into areas of vision, not on out-of-date design activity with demonstrable goals. documentation • Every scrum team has a customer who is • Work with multi-disciplinary teams to the keeper of the vision for a particular achieve vertical slice goals, whether this goal and establishes what “done” means. is a level, a mechanic or an in-game cinematic.
Product Owners and Design How are they structured? Customers Mechanics - AI Mechanics - Weapons
• A Product Owner is the communicator of the vision for the entire project. • Scrum answers the “How” question, but the “What” and “Why” needs to be answered by focused direction and vision. • Design Customers carry the vision on specific scrum goals (levels, mechanics, etc) Levels – Priority 1 Levels – Priority 2
6 How are they structured? Traditional Structure
The mall Producer level needs Jumping better needs a pacing Lead Lead tuning pass Tech Artist We’ll do a bit of AI We’ll do a Senior population. pass today. Senior Tech Artist
Junior Junior Tech Artist
Scrum Structure Scrum Structure
Producer
Senior Artist
Junior Artist Junior Tech
How do we plan? How do we plan?
• Focused “stories” replace • Focused “stories” replace design documentation design documentation • Focus Test • Focus Test • Focus on Action/Reaction • Let team determine path
Player Experience
7 How do we plan? Benefits: Iteration Loop
• Focused “stories” replace design documentation Finding • Focus Test Not yet. the fun? • Focus on Action/Reaction • Let team determine path What • Come Prepared Woo! about • But Respond to change now?
Iteration + Focus = Quality Benefits: Project Stability
Project Stability Project Stability • Time spent fighting the game is time where the product value decreases
8 Additional Benefits Challenges
• Project visibility to design team • Designers must cope with • Increased focus high rates of change. • Identify problems and • Short and long-term opportunities early • Responsibility increases • Members of the team • Management
Conclusion Additional Reading
• Product Value • Gamasutra.com: “Paper Burns” • Stability • These slides available on • Transparency AgileGameDevelopment.com • Iteration • Focused Goals • Find the fun as early as possible
The Art of Using SCRUM in Introduction the Creative Process
• Co-founder of Sammy Studios/ High Moon Studios • 14 years of experience in videogame industry • Manage and supervise: Emmanuel Valdez • Art department Chief Creative Officer • Audio department • Motion Capture • Cinematics department • Outsourcing
9 Agenda
• SCRUM in art production • How we use SCRUM • Challenges • Benefits • Evolution of SCRUM
?
Art Production is NOT Agile! Traditional Creative Goals
• Creativity- how unique? • Efficiency- how fast? • Consistency- is this the same game? • Effectiveness- is it good enough? • Quality- how good?
10 Pre-production Phase
• Smaller teams focus on creating and exploring new features Creativity • Iterative and exploratory This phase is usually 2 weeks • Concept art department • Identifying risks
Product Backlog Short Sprints (2 weeks)
• Incremental 2 Week Sprints • Focus • Velocity • Flexibility Reviews • Limiting resources
Feedback
Directors Publisher
Efficiency This phase is usually 2 weeks
11 Coordination and Communication
Tracking
Visibility Accountability
Transparency
Productivity SCRUM Team
Artists Consistency Technical Artists This phase is usually 2 weeks Game Designers Productivity Proprietary Tools Python Scripting
Asset Management Tools Enhancing middleware
The Agile Lead Artist Product Value
• Redefined roles and responsibilities • Focusing on “the big picture” • Focus on art creation and direction • Eliminates waste • No scheduling • Regular team reviews • Delegates authority • Prioritizing goals and • Mentoring and not managing ideas • Still executes on Art Director’s vision
12 Vertical Slice
Effectiveness This phase is usually 2 weeks
The Agile Artist Collaboration • Takes ownership
• Continuous input and feedback from Art Director and Product • Communicates and Owners coordinates with the other • Product Owner’s responsibilities: • Verifies work through leads disciplines • Aligns expectations with Art Director • References COS • Takes ownership over SCRUM teams goals and stories • Core of art production
The Agile Art Director
• Compliments vision of game design Quality • Not a manager This phase is usually 2 weeks • Focus on big picture • Sets the bar to achieve goals
13 Using SCRUM to Achieve SCRUM Quality Creative Goals
• Visual prototyping Creativity Effectiveness Iterative and incremental Vertic al slices Flexibility Redefining role of artists • Specialists Velocity Interdisciplinary coordination Focus • Adding polish tasks Quality • Detailed C.O.S. Efficiency Redefining Art Director Role Tracking dependencies Focus on the big picture Enhanced communication Specialists • Regular spot checks Visibility Polish Accountability Detailed conditions of satisfaction • Focus on product Transparency Regular spot checks value Consistency Redefining Lead Roles Focus on art direction Product value Productivity goals Regular reviews
Conclusion Thank You!
• Commitment • Email address: [email protected] • Visit us at GDC Expo! We are in Career Pavilion booth • Teamwork #1209 (Vivendi Games). • Talent • For company information online: www.highmoonstudios.com • Having people in the right key positions • Press inquiries: [email protected] • Facilities • Meelad Sadat • [email protected] • Don’t be afraid of changes • (310) 346-5325 • Faith • “The Art of Darkwatch”, www.designstudiopress.com
Working with Agile: A Production Perspective
• Steve Sargent • Executive Producer • High Moon Studios
14 Discussion Points My Production Philosophy:
• What the Producers do at HMS in • If the team is a rock band… relation to: • Daily Scrum • Sprints • Reviews & Planning • Releases • And a few other areas as well…
My Production Philosophy: Our Development Cycle
• …Production is the Road Crew 2-3 Weeks • (and the Producer the Band Manager) Sprint
2-3 Months
2-3 Years
• Our daily goal is to let the creative people focus most on being creative
Daily Scrum Producers @ Daily Scrum
• Producers are not usually Pigs • Producers are usually Scrum Master • 1. What did you do yesterday? • 2. What are you doing today? • 3. What are your impediments? • Track work left to do and new tasks • Remind team of its commitment
15 Daily Scrum: IMPEDIMENTS! Sprint Reviews
• ANYTHING hindering the team from 2-3 Weeks working optimally is an impediment Sprint • Most important question at daily Scrum • Not reporting impediments is a cardinal sin… • Removal of impediments is the most important job of the day for Production • Tracked using cards in the War-Room
Producers @ Sprint Reviews: Producers @ Sprint Reviews Backlog Maintenance
• Producers are not Product Owners • Producer “looks after” Product Backlog • Get the build ready for review • Backlog maintenance: Before the review • Haven’t blown it (yet!) • Team leads update & estimate Backlog • Backlog maintenance: Before planning • Review paperwork • Leads meet with Product Owner post review • Managing review day & prioritize the Product Backlog • Retrospectives: Mini Post-Mortems • Producer presents schedule and budgetary risks
Producers @ Sprint Planning Releases
• Schedule and facilitate planning 2-3 Weeks Sprint • Help create prioritized & clear User Stories 2-3 Months • Do not assign or estimate tasks • But do remind people not to over-commit • Create Burn-Down & pin up cards • External schedule & paperwork • Personnel movement
16 Producers and Releases Some other Production tasks
• Schedule the timeframes of Releases • General external liaison • Open space: Facilitate & follow up • Q.A. • Scores of action items that actually get • Outsourcing addressed! • Motion Capture • Adjust the long term plan • Dealing with H.R. (Calendar) • Crunch / O.T. (yes, this still happens)! • Longer term planning • Has yet to be enforced on current projects
After a couple of years… What we learned today
2-3 Weeks • The role Producers play at each stage of Sprint the Scrum process at High Moon Studios 2-3 Months • Producers do not micro-manage • Instead, we spend more time on mentorship, inter-team communication 2-3 Years and long term planning
What we learned today What next?
• We try to let the team spend as much of • I’ll answer any questions now its time working on the game content… • If you have any questions later, you can • …by handling (almost) everything contact me at: required to release a game except • [email protected] creating the content • [email protected] • You can also contact me through the IGDA Production mailing list
17 Agile Planning Imagine... • That you’re fed up with software development Mike Cohn as a career • And you decide to go into the landscaping business
• Your first job is moving this pile of rock from the front of my house to the back
How might you estimate this? My landscaping
• One way: 80 • Look at the pile of rock and estimate how many wheelbarrow loads it represents 60 • After an hour, see how many wheelbarrow loads you’ve moved then extrapolate the total duration 40
•I think that’s 80 wheelbarrow loads 20 •After an hour I’ve moved 20 loads •So, I’ll be done in a total of 4 hours Loads Wheelbarrow 0 0900 1000 1100 1200 1300 Time
••An iteration is a short, Relating the different planning constrainedconstrained periodperiod ofof timetime Velocity is the ••Typically 1-4 weeks levels amountamount ofof workwork Product Backlog Iteration Backlog plannedplanned oror As a frequent Code the UI 8 completedcompleted inin anan 3 flyer, I want to... 3 iteration.iteration. Write test fixture 6 As a frequent 2 2 5 Code middle tier 12 3 3 2 3 3 flyer, I want to... 2 Write tests 5 2 As a frequent 1 5 Automate tests 4 flyer, I want to... 3 2 3 As a frequent 2 2 4 2 2 2 2 2 4 flyer, I want to... “Yesterday I started As a frequent on the UI; I should 2 finish before the end A release typically comprises flyer, I want to... A release typically comprises of today.” more than one iteration
18 Story points Consider these two piles of work
• Probably the most commonly used estimating unit among agile teams today • Name is derived from agile teams commonly expressing requirements as “user stories” • Based on a combination of the size and complexity of the work • Unitless but numerically relevant estimates • A 10-point user story is expected to take twice as long as a 5-point user story What story point values might we put on these?
Three key advantages to Zoo points estimating in story points:
Assign “zoo 1. Forces the use of relative estimating points” to these • Studies have shown we’re better at this† animals Lion Kangaroo 2. Focuses us on estimating the size, not Rhinocerus the duration Bear • We derive duration empirically by seeing Giraffe how much we complete per iteration Gorilla 3. Puts estimates in units that we can add Hippopotamus together Tiger • Time based estimates are not additive † Lederer and Prasad, 1998. A Causal Model for Software Cost Estimating Error and Vicinanza et al., 1991. Software Effort Estimation: An Exploratory Study of Expert Performance.
Comparing apples to apples Planning poker for estimating • An iterative approach to estimating, loosely Product Backlog Iteration Backlog based on wideband Delphi As a frequent Code the UI 8 303 • Steps flyer, I want to... Write test fixture 6 1. Each estimator is given a deck of cards, each card As a frequent Code middle tier 12 has a valid estimate written on it 505 flyer, I want to... Write tests 5 2. Customer/Product owner reads a story and it’s discussed briefly As a frequent 505 Automate tests 4 3. Each estimator selects a card that’s his or her flyer, I want to... estimate As a frequent 202 4. Cards are turned over so all can see them flyer, I want to... “Yesterday I started on the UI; I should 5. Discuss differences (especially outliers) As a frequent 202 finish before the end 6. Re-estimate until estimates converge flyer, I want to... of today.”
19 Planning poker - an example Release planning Release Planning 8 13 20 3 5 2 Meeting 1
Estimator Round 1 Round 2 Release Plan Susan 3 5 Sprint 1 Sprint 2 Sprint 3 Sprints 4–7 Vadim 8 5 Ann 2 5 Chris 5 8
How’s my landscaping coming? Remember the different levels?
Product Backlog Iteration Backlog 80 As a frequent Code the UI 8 3 flyer, I want to... Write test fixture 6 60 As a frequent 5 Code middle tier 12 This is called a flyer, I want to... 40 We can track Write tests 5 burndown chart. As a frequent burndown5 at Automate tests 4 20 flyer, I want to...each level As a frequent
Wheelbarrow Loads 2 0 flyer, I want to... “Yesterday I started 0900 1000 1100 1200 1300 As a frequent on the UI; I should 2 flyer, I want to... finish before the end Time of today.”
An iteration burndown chart A release burndown chart
1,000 1,000
800 800
600 600
Hours 400
Story Points 400
200 200 0 0 1 2 3 4 5 6 7 5/6/02 5/13/02 5/20/02 5/24/02 4/29/02
20 Agile Practices
XP
Noel Llopis Senior Architect High Moon Studios
Agile Practices Automated Builds
• This should be common practice everywhere by now. • One-button full builds. No complex process. • Required foundation for other techniques.
Automated Builds Automated Tests
• Warning: • Unit tests: Test a single function or class. • Build needs to be robust and reliable. • Functional tests: Test whole program or • Bonus points for being fast. parts of it • Do both code and assets. Full game. • Fully automated. Run them on the build server at least once a day (we have them • Don't use if: running every couple of hours). • ??? Just do it. Never any reason not to. • Run them on your target platforms as well.
21 Automated Tests
Automated Tests Continuous Integration
• Warning: • Many small, very frequent check-ins • Need to be reliable enough to be trusted. (maybe every 3-5 minutes). • Must be run frequently and everybody help • As soon as any code is checked-in, build responsible for making sure they pass. is triggered. • People are notified right away of any failed • Don't use if: builds • Game is unstable or tests are expected to be broken. • Unit tests very useful keeping things • Lots and lots of cheap manpower (maybe). stable.
Continuous Integration
• House rule: Nobody leaves without making sure his last check-in resulted in a successful build. • The faster your builds, the better, so pay attention to logical and physical dependencies (feedback time for normal checkins is around 10-20 seconds). • Much slower for Unreal projects.
22 Continuous Integration Team Co-Location
• Warning: • Need to have automated builds and tests in place.
• Don't use if: • Main branch is unstable (fix it!) • Must work on many different branches.
Team Co-Location Collective Code Ownership
• Warning: • True collective code ownership, not "no • Avoid noisy environments. Move personal code ownership". areas away from work area. Encourage quiet • Really relies on unit tests and shared culture. knowledge.
• Don't use if: • You have distributed teams.
Collective Code Ownership Sustainable Pace
• Warning: • Avoid “no code ownership”! • All about maximizing long-term • People need to take pride in the codebase, even productivity. if they didn't write all of it. • Intense, focused workdays and then go home. • Don't use if: • You have no unit tests. • Team is only made out of specialists.
23 Sustainable Pace Sustainable Pace
Iterations • Warning: • Avoid punch card mentality. • Need tighter core hours to have team together. Intensity
• Don't use if: • Don't care for the long term or have an unlimited supply of junior developers.
Planning Implementation Alpha Beta Release Time
Pair Programming
Pair Programming Pair Programming
• Is output 2x of what one programmer can • Another benefit: Intensity and do? Not quite. concentration. You never have low • It's a bit less (around 1.7x), but also has moments, little breaks, or anything. You other major benefits: work 8 hours and you end up exhausted! • Much higher quality • Spread knowledge/philosophy • Team spirit • In the long term is a huge win
24 Pair Programming Test-Driven Development
• Warning: • Some people don't like it. • Not everything has to be done with pair programming. Write Write test code • Don't use if: • Have a small team made up of very senior programmers. Check in Check in Refactor
TDD Benefits: Simplicity, Modularity TDD Benefits: Safety Net
TDD Benefits: Documentation TDD != Unit tests TDD != Testing strategy
TDD == Development technique
25 I'd like to use TDD but... Test-Driven Development
• Warning: • You'll take an initial productivity hit. • Hard and discouraging with a bad codebase. • Everybody must be on board.
• Don't use if: • You're writing throwaway code and nothing depends on it. • Can't afford to take the initial hit.
Outline
• Background of Epic Agile Development in Practice at • The Problem: Epic’s Unique Challenges Epic Games • The Solution: Epic’s Flavor of Scrum Dr. Michael Capps, President • The Fallout: Experience So Far Daniel Vogel, Engine Programmer
Epic Games Unreal Engine History
• Founded 16 years ago by • In active development for 12 years CEO Tim Sweeney • Licensed to external developers since 1997 • About 80 employees (18 engine devs) • More than 50 games have used Unreal Engine • Early success with Jill of the • Unreal Engine 1 (1997-2000) Jungle and Epic Pinball • PC First-Person Shooters • Multiple hits in the Unreal and Unreal Tournament series • Deus Ex, Rune, Undying, Harry Potter • Recently shipped Gears of • Unreal Engine 2 (2000-2004) War, the fastest and highest • PC, Xbox, PS2, Mac, Linux; mostly First-Person Shooters selling 360 title • Splinter Cell, America’s Army, Lineage II, Rainbow 6
26 Unreal Engine 3 UE3 Games in Development
• Unreal Engine 3 (2006-??) • A few licensees: • Real Time Worlds • Atari • Sony Online • next-generation consoles and PCs • Bioware • Square Enix • supports any game type, even massive • Disney / BVG • Take Two multiplayer and fighting games • Electronic Arts • THQ • work began 4.5 years and 55 man years ago • EIDOS • Ubisoft • Gearbox Software • US Army • Game Developer Magazine award for best • Microsoft • Vivendi engine/middleware 3 years running • Midway • Webzen • NC Soft
Internal and external Outline customers
• Background of Epic • Licensees whose game schedules, and • The Problem: Epic’s Unique Challenges thus their entire business, depends on our engine • The Solution: Epic’s Flavor of Scrum • Very large range of needs • The Fallout: Experience So Far • From Wheelman to Mortal Kombat – and that’s one company • Multiple internal game teams with competing needs
Decision Making Quality Assured Builds • Prioritizing tasks is quite complex when balancing internal teams, external teams, • We give licensees up-to-the-minute competition, marketing needs, etc. source code access • ‘agile’ in that we would switch tasks quickly as • Which means they can easily access broken priorities changed code • Task backlog was so large that we had a “do it • Thus we would regularly checkpoint with now or it’ll never get done” attitude that voided ‘Quality Assured’ builds which go through prioritization schemes full regression testing • Internal customers took advantage of this by • Our goal was to do this every 1-2 weeks direct discussion with programmers • In practice there was a negative feedback loop that made delivery vary from 2-8 weeks
27 Internal Communication External Communication • Licensees get a regular report from bug database to see which bugs have been closed • Bug tasks stored in a bug database • Licensees given list of large-scale ‘planned • Features listed on an internal wiki page features’ with months/quarters attached with estimates • Many tasks slipped significantly – usually due not to • Weekly meeting that would run an hour bad estimation, but to unplanned tasks taking or more for complex topics precedence • Daily status emails from each engineer • No way to learn timeframe/priority of smaller that would often spark discussion. Read tasks without direct inquiry by game team coding leads / producers.
Outline Our brand of scrum
• Background of Epic • 20 work days per sprint • The Problem: Epic’s Unique Challenges • High specialization prevents dynamic • The Solution: Epic’s Flavor of Scrum task shuffling • The Fallout: Experience So Far • Lots of unscheduled time (28% average, up to 40% max) • Code lockdown at end for QA release • Highly favor release date over feature set, punting tasks to next sprint but never dropping entirely
Scrum Tools Outline
• Sprint backlog managed via Excel and • Background of Epic TestTrack Pro (TTP) • The Problem: Epic’s Unique Challenges • Project backlog managed via TTP • The Solution: Epic’s Flavor of Scrum • Coarse view of project backlog available • The Fallout: Experience So Far to licensees via road map UDN page • Future sprint Excel sheet contains all major roadmap items in combination with high priority smaller tasks from project backlog
28 Pros Cons
• Convenient communication with • Treadmill effect licensees • Increased minimum time between QA • Progress tracking (motivating) builds, in practice increased frequency • Easier for us to schedule larger tasks! • Increased minimum latency for non- • Gives name to what we’ve been doing blocking issues • Reduces randomization, increasing • Code locking has impact on game teams efficiency • Crash report spike after lock is lifted • Game teams involvement with • No penalty for over/ under- estimating scheduling due to punting • Daily status emails
Questions?
• Thanks for staying until the end of the day! Any Questions? Clinton Keith – Closing Comments • [Yes, we’re hiring, visit our job fair booth or contact [email protected]]
• More info at: • www.epicgames.com • www.unrealtechnology.com
Benefits of agile Closing Comments • Improved… • Productivity • Especially for leads • Reliability of build • Benefits of agile • Quality of game • Morale • What still needs to be addressed • Ownership • Where to go from here • Team work • Communication • Enables low-cost executive awareness • Which means you don’t have to work hard to find out what’s going on • Makes us constantly see and think about what we are doing. Encourages common sense.
29 Going forward…. Where to go for more info? • Production is not as agile • Minimum content required to ship • “Lean” principles have value • www.agilegamedevelopment.com • Large teams need more leadership • www.gamesfromwithin.com • Self-organization breaks down • Ownership/commitment much more difficult • www.mountaingoatsoftware.com • Vertical slices are hard • www.epicgames.com • Debt sneaks in • Releases can be mini-waterfalls • www.unrealtechnology.com • These cause crunch • Agile makes us constantly see and think about what we are doing so we can make common sense decisions.
Thanks for attending!
30