Vernaq
Login Sign Up

Things Developers Say

Public
by Vernaq • 109 items

A translation guide to the phrases developers use when talking about code, bugs, deployments and the mysterious ways software behaves in production.

Items

Alert fatigue
Getting so many alerts that nobody reacts properly anymore.
Auth
The part that proves who you are and breaks at the worst possible time.
Authz
Checking what an authenticated user is allowed to do.
Backfill
Populate old data after adding something new.
Backward compatible
The new version still works with the old way of doing things.
Bikeshedding
Spending too much time debating something small and unimportant.
Blue-green deployment
A deployment strategy designed to reduce risk when releasing.
Boilerplate
Repetitive code required by a framework.
Bottleneck
The slowest part of the system holding everything else back.
Breaking change
An update that forces everyone else to change their code too.
Build
The compiled version of the application.
Cache invalidation
One of the classic ways software becomes surprisingly difficult.
Caching
Saving results so you don’t have to compute or fetch them again.
Canary release
Release to a small group first and see if anything explodes.
Cannot reproduce
The bug exists somewhere, just not where you’re looking.
CI pipeline
Automated tests and checks run on every change.
CLI
A command-line interface for developers who prefer terminals to buttons.
Code review
Another developer checking your code.
Compile error
A problem detected before the program runs.
Connection pool
A managed set of reusable database or service connections.
Contract test
A test that makes sure two systems still agree on how they communicate.
Cron job
A scheduled task that quietly runs in the background until it fails at 3am.
Deadlock
Two processes waiting forever on each other.
Dependency
A library the project relies on.
Deploy
Release new code to production.
Deprecated
Still there for now, but living on borrowed time.
Edge case
A rare scenario nobody originally thought about.
End-to-end test
A test that simulates real user behaviour across the full system.
Error budget
How much unreliability the team can tolerate before slowing down releases.
Event-driven
A design where systems respond to events instead of direct calls.
Fallback
The backup behaviour used when the preferred option fails.
Feature flag
A switch that lets you turn functionality on or off without redeploying.
Flaky test
A test that sometimes passes and sometimes fails for no good reason.
Formatting
Making the code look the way the team expects.
Full table scan
When the database has to read everything because you made it sad.
Graceful degradation
Failing in a controlled and less awful way.
Green build
Everything passed and nobody is panicking.
Hack
A quick and dirty solution.
Happy path
The ideal user flow where everything goes right.
Heisenbug
A bug that disappears or changes when you try to inspect it.
Hotfix
A rushed patch to fix something broken in production.
Idempotent
Safe to run multiple times without changing the outcome.
Incident
A production problem important enough to get everyone’s attention.
Index
A database structure that makes lookups faster when used properly.
Integration test
A test to make sure separate pieces work together.
Latency
The time delay between request and response.
LGTM
Looks good to me.
Linting
Automatic checks for style and obvious mistakes.
Logging
Writing down what the system is doing so future-you can investigate it.
Memory leak
Memory that is allocated but never released.
Next »
Vernaq is in beta — we’d love your feedback. Send Feedback