- the business side of things is completely terrible
- executives at big companies can totally ruin your company, and not care, when they acquire you
- just went ahead and rewrote it, whee
- identify the target
- don’t let people know what you’re doing at first or try to convince people beforehand
- 1. show, don’t tell
- 2. forgiveness, not permission
- people get locked into the JVM mindset
- 100% political problem
- call everything a proof of concept, even if it’s aimed at production
- once it works, be very careful
- demo to the most powerful person you can get a meeting with
- air cover: trade something that works for political protection, in big companies working software is currency
- two lines of people:
- people who want to take credit for your work
- people who want to take you down because your work threatens their jobs
- who is going to benefit? find them and create alliances with them
- want it to look like a fait accompli
- once your thing is in production, you need to take care of:
- operations: you have to train your org or deskill running it; they used terraform
- scripting setup
- stress test
- monitoring and alerts and oncall
- runbooks
- recruiting:
- internal recruiting: sell OTP via the remote shell and tracing and inspection
- don’t require OTP experience; rules out too many people
- hire everywhere but the bay area
- scale your team, make yourself redundant
- operations: you have to train your org or deskill running it; they used terraform
- second JVM system you replace is easier
Created
March 2, 2019 01:35
-
-
Save evanmcc/eb63cfa71c76694072d265af7df304b6 to your computer and use it in GitHub Desktop.
talk on the politics of introducing beam services at jvm shops.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Real-life example of replacing JVM with BEAM: JVM strugles and the BEAM