Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Simplifying Software Estimation
Search
Pratul Kalia
July 24, 2020
Programming
1
210
Simplifying Software Estimation
Presented at GDG MAD in July 2020.
Pratul Kalia
July 24, 2020
Tweet
Share
More Decks by Pratul Kalia
See All by Pratul Kalia
The special case of Mobile DevOps
pratul
2
23
Reldex: measuring the effectiveness of your app release process
pratul
0
11
Stop the security theater!
pratul
0
250
Effective and efficient mobile engineering
pratul
0
160
Designing future-proof Android applications
pratul
0
150
Android - an introduction for developers
pratul
3
220
Semantic Content Repositories
pratul
1
120
How To Become A Hacker
pratul
3
300
Other Decks in Programming
See All in Programming
Vibe Codingをせずに Clineを使っている
watany
17
6.3k
Golangci-lint v2爆誕: 君たちはどうすべきか
logica0419
1
120
ウォンテッドリーの「ココロオドル」モバイル開発 / Wantedly's "kokoro odoru" mobile development
kubode
1
140
Sharing features among Android applications: experience feedback
jbvincey
0
110
「理解」を重視したAI活用開発
fast_doctor
0
110
Make Parsers Compatible Using Automata Learning
makenowjust
1
5.1k
State of Namespace
tagomoris
4
1.8k
PHPで書いたAPIをGoに書き換えてみた 〜パフォーマンス改善の可能性を探る実験レポート〜
koguuum
0
170
MCP調べてみました! / Exploring MCP
uhzz
2
2.3k
小田原でみんなで一句詠みたいな #phpcon_odawara
stefafafan
0
340
Fiber Scheduler vs. General-Purpose Parallel Client
hayaokimura
1
100
Road to RubyKaigi: Making Tinny Chiptunes with Ruby
makicamel
4
380
Featured
See All Featured
[RailsConf 2023] Rails as a piece of cake
palkan
54
5.4k
Mobile First: as difficult as doing things right
swwweet
223
9.6k
How GitHub (no longer) Works
holman
314
140k
Stop Working from a Prison Cell
hatefulcrawdad
268
20k
Embracing the Ebb and Flow
colly
85
4.6k
Producing Creativity
orderedlist
PRO
344
40k
Done Done
chrislema
183
16k
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
135
33k
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
160
15k
ピンチをチャンスに:未来をつくるプロダクトロードマップ #pmconf2020
aki_iinuma
119
51k
The MySQL Ecosystem @ GitHub 2015
samlambert
251
12k
Understanding Cognitive Biases in Performance Measurement
bluesmoon
29
1.6k
Transcript
Pratul Kalia @PRXTL • OBVIOUS.IN SIMPLIFYING SOFTWARE ESTIMATION
SOFTWARE ESTIMATION STEVE MCCONNELL 2006, MICROSOFT PRESS
WHAT IS AN ESTIMATE?
LET’S DRIVE FROM ONE END OF THE CITY TO ANOTHER!
“How much time will it take to finish all these things?”
Estimates always have a probability.
Targets are not estimates!
THE TRAIN LEAVES AT 4.40 PM! “Here are a bunch
of things we have to get done by X”
Are you chasing a target or an estimate?
PRESSURE & BAD ESTIMATES
— ALL ENGINEERS “Okay, that is a bit of work
but I’m sure I can do it in 3 days.”
Stop putting yourself in unreasonable situations.
There are no prizes for getting there fast!
— PARKINSON’S LAW “Work will expand to fill all available
time.”
(that is… if teams have lots of extra time, more
work will magically get created.)
— GOLDRATT’S STUDENT SYNDROME “Given lots of time, students will
delay studying until the last possible moment”
Underestimation is worse than Overestimation! Stastically reduced chance of on-time
completion.
Engineers are too optimistic already.
OTHER SIDE EFFECTS… Poor technical foundation, worse in the long
run.
OTHER SIDE EFFECTS… Unproductive, emotionally draining behaviours.
OTHER SIDE EFFECTS… - more status meetings! - frequent re-estimation!
- interim releases!
OTHER SIDE EFFECTS… - meetings to cut scope! - fix
bugs arising from hacks!
None
THE CONE OF UNCERTAINTY
None
“As the project progresses, estimates become more accurate.”
(yes that’s quite , isn’t it.)
None
It is easily possible to do worse… but not better.
The Cone does not narrow itself.
None
None
Stop doing first-number-in-your-head estimates! It sets the wrong expectation!
⏱
Even a 15-minute estimate is exponentially better than an instant
one.
INCLUDE EVERYTHING. Some requirements are stated. Others are implied.
INCLUDE EVERYTHING. Deployment setup Maintaining build scripts Code reviews
INCLUDE EVERYTHING… Writing documentation Onboarding new team members
INCLUDE EVERYTHING! Creating test data Upgrading dependencies
FACTORS THAT INFLUENCE ESTIMATES
Size of the software! The biggest contributor to project effort
and schedule.
As size increases, effort goes up exponentially. Exponentially. Not linearly.
Programmer capability. Time constraints. Team continuity. Required reliability.
COCOMO-2 software estimation model.
IMPROVING YOUR ESTIMATES
I hope you use story points of some type.
STORY POINTS. Mark of complexity They don’t (necessarily) represent time!
STORY POINTS. (for e.g. A team using 1/2/3 points for
a project.)
STORY POINTS. A junior engineer takes 5 days to finish
a 3-point task.
STORY POINTS. A senior engineer takes 2 days to finish
a 3-point task.
The time taken is different. But the complexity is similar!
NOT COMPLEX… Create a “lint step” on the CI server
NOT COMPLEX… Add “patient has already visited” option to Overdue
list
COMPLEX? Allow Blood Pressure measurements to be deleted
The Law of Large Numbers
“If you create one Big Estimate, your error(s) will either
be completely on the high side, or the low side”
“but if you create multiple Small Estimates, some will be
on high side, some on the low side.”
“… so the errors will cancel each other out, to
some extent.”
COMPLEX? Allow Blood Pressure measurements to be deleted
None
ALSO COMPLEX Sync Protocol Drugs across all facilities and display
in the app
None
Some tricks!
TRICKS! Do a detailed task breakdown: spend some time!
TRICKS! Think of the “worst case” while estimating.
TRICKS! Compare different stories of similar complexity.
TRICKS! Use a project management tool that understands “velocity”.
TRICKS! Engineers should break down tasks and estimate them.
TRICKS! Look at past data for your team, for similar
type of work.
TRICKS! Measure yourself privately
NEGOTIATIONS & AGREEMENTS
An estimate is a conversation, and a negotiation.
Engineers “own” the estimate… but Management “owns” the target.
The goal of estimation is NOT pinpoint accuracy!
People want to understand VALUE in comparison to EFFORT.
Do t-shirt-sizing with all stakeholders.
VALUE EFFORT A: S M B: M S C: S
L D: XL M
Separate people from the problem.
Focus on interests, not on positions.
Come up with options that benefit everyone.
Either everyone wins, or everyone loses.
GO FORTH AND ESTIMATE!