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
The Scientist & The Engineer
Search
Andrew Godwin
February 07, 2020
Programming
860
1
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
The Scientist & The Engineer
My keynote from PyCon Colombia 2020.
Andrew Godwin
February 07, 2020
More Decks by Andrew Godwin
See All by Andrew Godwin
Reconciling Everything
andrewgodwin
1
400
Django Through The Years
andrewgodwin
0
320
Writing Maintainable Software At Scale
andrewgodwin
0
520
A Newcomer's Guide To Airflow's Architecture
andrewgodwin
0
420
Async, Python, and the Future
andrewgodwin
2
740
How To Break Django: With Async
andrewgodwin
1
830
Taking Django's ORM Async
andrewgodwin
0
850
The Long Road To Asynchrony
andrewgodwin
0
770
Pioneering Real-Time
andrewgodwin
0
530
Other Decks in Programming
See All in Programming
Hunting Vulnerabilities in Symfony with LLMs
vinceamstoutz
0
570
Even G2とAWSで推しのエージェントを召喚しよう!
har1101
1
140
ローカルLLMでどこまでコードが書けるか -縮小版 / How much code can be written on a local LLM Shortened
kishida
2
170
AIエージェントで 変わるAndroid開発環境
takahirom
1
330
鹿野さんに聞く!『TypeScriptコードレシピ集』で磨く実践力
tonkotsuboy_com
4
990
【やさしく解説 設計編 #0】DDDのコード、読めるのに分からない人へ
panda728
PRO
1
190
「AIで開発し、AIを届ける」をEvalでつなぐ 〜AIネイティブに始めるプロダクト開発の実践〜 / Connecting "Develop with AI, deliver AI" with Eval
rkaga
4
5.6k
アルゴリズムは何を圧縮しているのか ─ Haskell から育った「圧縮代数」というメンタルモデル
naoya
11
1.4k
AI がコードを書く時代における新卒エンジニアの仕事風景 (2026) / New Graduate Engineers in the Era of AI Coding (2026)
sushichan044
0
200
過去最大のMCPアップデート! 2026-07-28 RC版の謎に迫る
licux
6
440
AI時代のUIはどこへ行く?その2!
yusukebe
22
7.7k
Haskell/Servantを通してWebミドルウェアを捉え直す
pizzacat83
0
370
Featured
See All Featured
Have SEOs Ruined the Internet? - User Awareness of SEO in 2025
akashhashmi
0
380
Practical Orchestrator
shlominoach
191
11k
The B2B funnel & how to create a winning content strategy
katarinadahlin
PRO
1
410
How to Build an AI Search Optimization Roadmap - Criteria and Steps to Take #SEOIRL
aleyda
1
2.1k
The #1 spot is gone: here's how to win anyway
tamaranovitovic
3
1.1k
Navigating Algorithm Shifts & AI Overviews - #SMXNext
aleyda
1
1.3k
Winning Ecommerce Organic Search in an AI Era - #searchnstuff2025
aleyda
1
2.1k
RailsConf 2023
tenderlove
30
1.5k
A better future with KSS
kneath
240
18k
<Decoding/> the Language of Devs - We Love SEO 2024
nikkihalliwell
1
270
Product Roadmaps are Hard
iamctodd
55
12k
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
35
2.5k
Transcript
THE SCIENTIST ANDREW GODWIN // @andrewgodwin THE ENGINEER AND
Hi, I’m Andrew Godwin
I call myself an "engineer"
But I studied Computer Science
I knew programming going in. But this was different.
“ Edsger Dijkstra Computer science is no more about computers
than astronomy is about telescopes.
Computer Science is idealistic
"Turing-complete"
We first want to prove it is possible
async def sleep_sort(numbers): async def sleep_add(number): await asyncio.sleep(number) result.append(number) result
= [] await asyncio.wait([ sleep_add(number) for number in numbers ]) return result
We can prove that programs are correct
def is_two(number): return number == 2
>>> import __builtin__ >>> __builtin__.True = False >>> True False
>>> True == False True (Thankfully, only works in Python 2!)
"Communicating Sequential Processes", C. A. R Hoare
contains :: (Ord a) => (Tree a) -> a ->
Bool contains Nil _ = False contains (Node t1 v t2) x | x == v = True | x < v = contains t1 x | x > v = contains t2 x
A perfect world, of perfect programs
“ Benjamin Brewster In theory there is no difference between
theory and practice. In practice there is.
Software Engineering takes shortcuts
When is it safe to take a risk?
None
Software is faster and cheaper to change
The real world is a nasty place
How often does a cosmic ray affect RAM? 36 hours
(for 16GB; "SEU at Ground Level", Eugene Normand) How long can an unpowered SSD keep data? 3 weeks - 1 year (Enterprise at 40ºC, Client at 30ºC; Intel/JEDEC) Does quantum tunneling affect CPUs? Continuously!
Always design for failure
How do we reason about software?
Input Process Output
async def disable_alarm(): ... async def open_door(): ... await asyncio.wait(
[disable_alarm(), open_door()] )
Input Rendering Output Validation Reporting Storage Processing Logging Cache
Django has 250,000 lines of code
An Airbus A380 has 4,000,000 parts
A new car has 100,000,000 lines of code
How are we supposed to handle this?
Abstract, verify and forget
Abstract Define a contract - types, behaviour, exceptions Verify Write
tests to keep the contract valid Forget Work with the contract, not the fine details
You have to learn to forget
Engineering is communication
“ Grace Hopper A ship in port is safe, but
that's not what ships are built for.
Build to expect growth
Build to expect failure
Everyone can build near-perfect software
Know when not to!
Python lets you do both.
Types Start without types, progressively add with mypy Async Make
it work synchronously first. Add async later. Speed Write slow, understandable code. Test it. Then improve it.
Scientists observe and question
Engineers build and invent
All software has consequences
Be the scientist and the engineer
Gracias. Andrew Godwin @andrewgodwin // aeracode.org