How to burn out developers and alienate contributors

About me: oli-obk

Note:

obviously biased

survivorship bias

tell me where I can do better

The plan

• a bit about research
• a bit about open source
• drawing parallels
• the future

Note:

lots of anecdotes, starting out a bit dark

Research

1. Have idea
2. Find Prof who likes idea
3. Do PhD
4. Finish PhD
5. Profit!?!

Not so fast

teaching
exams
grading
random PR campaigns for funding
find funding for yourself
mentor young PhDs
watch out for rival profs' proxy warring
write papers for your prof
figure out what you were working on a month ago
actually work on your topic

Note:

don't read all that

I love teaching, just not how it's often done

Anyone heard of burnout?

Spoiler: Not in academia

Note:

(outside of psych)

[redacted]

Note:

had a bunch of quotes from awful ppl here

not worth repeating.

be aware that science isn't the pretty ivory tower

Anyway…

Not everyone finishes their thesis

Drop out rate (Germany)

Never trust data you didn't falsify yourself

Note:

Number of unreported cases

unregistered until a year or two in

Summary

Research has lots of exploitation

… among other problems

Open Source

Note:

earnest OS contrib since 2016

Note:

dove into rustc

stuck around

Quick Reminder

survivorship bias

1. Have idea
2. implement
3. open PR
4. get review
5. address review
6. merge

Not so fast

• reviewing costs time
• 8h work per day
• 6h effective per day [1]
• why is my PR not getting reviewed?

[1] https://www.theguardian.com/technology/2019/nov/04/microsoft-japan-four-day-work-week-productivity

Note:

open source after work hours

already drained

weekend?

Solution:

Become a reviewer

Note:

save the other maintainers some time

You guessed it

Not. So. Fast!

Note:

I joke, but…

actively reviewed projects attract

People come …

… and go

Note:

Sometimes they go because of you

Note:

subjective experience

data exists, but… hard to analyze

Parallels

Let's fix it

Note:

joke: ask 10 ppl, get 12 opinions

all boils down to 3 things

Note:

three components

• money
• time
• resources

no particular order

Resources

• Reduce compile-time/test-time
• Speed up CI
• Personal hardware

Time

• Minimize your tests
• Bisect regressions
• Accept maintainer priorities
• Deadlines are a lie
• Real vacations or sabbaticals

Note:

deadlines and lighthouse projects can quickly make some people throw the towel

encourage ppl to really leave for a while to recharge

Money

• Sponsor your maintainers
• Hire your maintainers
• Pay people to become maintainers

Progress in Rust

• ~10 FTE on Compiler hired
• Library maintainers hired
• Paid on-call crates.io team
• WIP: ssh dev-desktop.rust-lang.org
• Sponsoring developers…

Note:

Many hired devs don't get 100% though

No, this URL doesn't exist yet

Sponsoring

Note:

Shout out to sponsors.

"no particular order", just the order I found them

looking through some Rustacean's gh sponsors

Foundation members working on eliminating red tape

Hiring is easier than sponsoring

Personal note

Note:

connecting the dots of this talk with my life

2016-2020: PhD + Rust

• Working 10+ hours/day
• Working weekends

Note:

Was already planning my Rust exodus

Feburary: Working at aws

• BigCo trainings
  Image Not Showing Possible Reasons

  • The image file may be corrupted
  • The server hosting the image is unavailable
  • The image path is incorrect
  • The image format is not supported

  Learn More →
• Work on rustc full time
  Image Not Showing Possible Reasons

  • The image file may be corrupted
  • The server hosting the image is unavailable
  • The image path is incorrect
  • The image format is not supported

  Learn More →
• Weekends 🎉

Note:

Same work as before, but paid

April: Parental leave

July: Part time

Note:

Practiced this presentation over the weekend

Summary

• Pay devs
• Respect devs' time
• Simplify devs' lives
• Rust Foundation!!!

Note:

Foundation's mission is to support maintainers