Tech glossary
Software words, in plain English.
134 terms managers hear every week, each explained in one or two sentences. For anything not listed, paste it into Plainlee.
A
- A/B test
- Showing two versions of something to different users to see which one works better, based on real results instead of opinions.
- Acceptance criteria
- The checklist that says when a piece of work counts as done, agreed before the work starts.
- Agile
- A way of working in short cycles, delivering small pieces often and adjusting the plan as you learn.
- API
- A doorway that lets two pieces of software talk to each other, such as your app and a payment provider.
- API key
- A secret code that lets one piece of software use another's service. It should never be shared in public.
- Authentication
- Checking who someone is, usually with a password, a code or a company login.
- Checking what someone is allowed to do once they are logged in, such as view only or edit.
B
- Backend
- The behind-the-scenes part of the product that stores data and does the work. Users never see it directly.
- Backlog
- The list of work the team has not started yet, usually in priority order.
- Beta
- An early version shared with some users to find problems before the full launch.
- Blocked
- Work cannot move forward until something else happens first, like a decision, access or another team's work.
- Branch
- A separate copy of the code where a developer makes changes without affecting everyone else.
- Breaking change
- An update that stops older things from working until they are updated too.
- Bug
- Something in the software that does not work the way it should.
- Build
- The step that turns the team's code into a version of the product that can actually run.
C
- Cache
- A saved copy of data kept close at hand so the product loads faster. Old cached data can make changes look like they did not work.
- Changelog
- A running list of what changed in each version of the product.
- CI/CD
- Automatic checks and releases. Every code change is tested (CI) and can be shipped to users automatically (CD).
- Cloud
- Renting computers and storage from a provider like AWS, Google Cloud or Azure instead of owning servers.
- Code review
- Another developer reads a change before it is accepted, to catch mistakes early.
- Commit
- A saved snapshot of a developer's code changes, with a short note about what changed.
- Container
- A packaged-up version of the app with everything it needs, so it runs the same way on any server.
- Cron job
- A task the system runs automatically on a schedule, such as sending a report every morning.
D
- Database migration
- A planned change to how data is stored. It often needs care because real customer data is involved.
- Debugging
- Finding out why something is broken. It can take minutes or days, depending on how hard the cause is to find.
- Definition of done
- The team's shared standard for when work is truly finished, for example tested, reviewed and documented.
- Dependency
- Something the work relies on, such as another team, a tool or a piece of outside code.
- Deploy
- Putting a new version of the software live for users.
- DNS
- The internet's address book. It connects a web address like your-company.com to the server that runs it.
- Downtime
- Time when the product is not working or not reachable for users.
E
- Edge case
- An unusual situation that only happens sometimes, such as a customer with no last name.
- Encryption
- Scrambling data so only the right people or systems can read it.
- Endpoint
- One specific address in an API that does one job, such as getting a customer's orders.
- Environment
- A separate copy of the product used for one purpose: development for building, staging for testing, production for real users.
- Epic
- A large piece of work broken down into smaller tasks or stories.
- Estimate
- The team's best guess of how long work will take. It is a range, not a promise.
F
- Feature flag
- A switch that lets the team ship new code but keep it hidden until someone turns it on.
- Flaky test
- An automatic check that sometimes passes and sometimes fails without any code change, which slows the team down.
- Fork
- A copy of a project that goes its own way, often to make changes without touching the original.
- Framework
- A ready-made foundation developers build on so they do not start from zero, such as React or Laravel.
- Frontend
- The part of the product people see and click: screens, buttons and forms.
G
- GDPR
- European data protection law. It sets rules for collecting, storing and deleting people's personal data.
- Git
- The tool most teams use to track every change to the code and who made it.
- GitHub
- A website where teams store their code, review changes and track work. GitLab and Bitbucket are similar.
- Grooming
- A meeting to tidy the backlog: clarify upcoming work, estimate it and put it in order. Also called refinement.
H
- Hardcoded
- A value written directly into the code, so changing it later needs a developer instead of a settings page.
- Hosting
- The service that keeps the product running on servers and reachable on the internet.
- Hotfix
- An urgent fix released quickly to solve a serious live problem.
- HTTPS
- The secure version of a web connection, shown by the padlock in the browser. Data travels encrypted.
I
- Incident
- A live problem that affects users, such as an outage or slowness. Often graded by severity, like P1 for the most serious.
- Infrastructure
- The servers, networks and services the product runs on. Often shortened to infra.
- Integration
- Connecting the product to another tool or service so they share data, such as a CRM or a payment provider.
- Integration test
- An automatic check that several parts of the product work correctly together.
- Iteration
- One round of building, checking and improving. Each iteration gets the product a bit closer to the goal.
J
- JavaScript
- The main programming language of websites. It makes pages interactive.
- Jira
- A popular tool for tracking tickets, sprints and project progress.
- JSON
- A common, text-based format for sending data between systems. It looks like labelled lists of values.
K
- Kanban
- A way to manage work on a board with columns like To do, In progress and Done, focused on steady flow.
- Kickoff
- The first meeting of a project, where goals, scope, roles and timelines are agreed.
- Kubernetes
- A system that runs and manages many containers across servers. Often shortened to K8s.
L
- Latency
- How long the product takes to respond. Higher latency means a slower experience.
- Legacy code
- Older code that still runs the product but is harder to change safely.
- Load balancer
- A traffic controller that spreads visitors across several servers so no single one is overloaded.
- Localhost
- A developer's own computer. If something works on localhost, it has not been tested on a real server yet.
- Logs
- Records the system keeps of what happened and when. Developers read them to find the cause of problems.
M
- Merge
- Combining one developer's changes into the main version of the code.
- Merge conflict
- Two developers changed the same part of the code, so someone has to decide how to combine them. Usually hours, sometimes a day.
- Microservices
- Building the product as many small, separate services instead of one big program.
- Milestone
- An important checkpoint in a project, such as a beta launch or a key feature being finished.
- Monitoring
- Tools that watch the live product and alert the team when something goes wrong.
- MVP
- Minimum viable product: the smallest version that is useful enough to put in front of real users.
N
- No-code
- Tools that let people build apps or automations without writing code, such as Zapier or Webflow.
- Node.js
- A popular way to run JavaScript on servers, often used for a product's backend.
- Null
- A value that means nothing is there. Unexpected nulls are a common cause of errors.
O
- OAuth
- The standard behind Sign in with Google or Microsoft. It lets apps get access without seeing your password.
- Onboarding
- The first steps that help a new user or team member get started.
- Open source
- Software whose code anyone can see, use and improve, usually for free.
- Outage
- A period when the product, or part of it, stops working for users.
P
- Patch
- A small update that fixes a problem, often a bug or a security issue.
- Penetration test
- A security test where experts try to break in, the way an attacker would, to find weak spots.
- Product owner
- The person who decides what the team builds next and in what order.
- Production
- The live product that real customers use. Often shortened to prod.
- Prototype
- An early, rough model of an idea used to test it before building the real thing.
- Pull request
- A request to add a developer's change to the main code, where it is reviewed first. Also called a PR or merge request.
Q
- QA
- Quality assurance: testing the product to find problems before users do.
- Query
- A request to a database for specific information, such as all orders from last week.
- Queue
- A waiting line for tasks the system handles one by one, such as sending emails. A long queue means delays.
R
- Rate limit
- A cap on how many requests a service accepts in a period. Hitting it slows things down or causes errors.
- Rebase
- Updating a developer's branch with the latest main code so the change fits cleanly. Routine, but it can surface conflicts.
- Refactor
- Rewriting code to make it cleaner without changing what it does. It makes future work faster and safer.
- Regression
- Something that used to work and broke after a recent change.
- Release
- A bundle of changes shipped to users together.
- Repository
- The place where a project's code and its history are stored. Often shortened to repo.
- Requirements
- A description of what the product needs to do, agreed before building.
- Retrospective
- A meeting at the end of a sprint to discuss what went well and what to improve. Often called a retro.
- Roadmap
- The plan of what the team will build over the coming months, and roughly when.
- Rollback
- Undoing a release and going back to the previous version, usually because something went wrong.
S
- SaaS
- Software as a service: software you use online with a subscription instead of installing it.
- Sandbox
- A safe, separate place to try things without affecting real data or users.
- Scalability
- How well the product handles growth, such as many more users, without slowing down or breaking.
- Scope creep
- The work keeps growing beyond what was agreed, which pushes dates back.
- SDK
- A ready-made toolkit a company provides so developers can connect to its service faster.
- Server
- A computer that runs the product and answers requests from users' browsers and apps.
- SLA
- Service level agreement: a promise about uptime or response times, often with penalties if missed.
- Spike
- A short, time-boxed research task to answer a technical question before committing to an approach.
- Sprint
- A fixed period of work, usually one or two weeks, with an agreed set of goals.
- SSO
- Single sign-on: people log in once with their company account instead of a separate password.
- Staging
- A test copy of the product that matches the live one, used for checks before release.
- Stand-up
- A short daily team meeting about progress, plans and blockers. Usually 15 minutes.
- Story points
- A rough measure of how much effort work takes, compared with other work. Not hours.
T
- Tech stack
- The set of technologies a product is built with, such as its languages, frameworks and databases.
- Technical debt
- Shortcuts taken earlier that now slow down new work until they are cleaned up.
- Third-party
- A tool or service made by another company that the product relies on.
- Ticket
- One recorded task or issue in a tool like Jira, with a description, owner and status.
- Timeout
- When a system waits too long for an answer and gives up, usually showing an error.
- Token
- A digital pass that proves who you are or what you can access, often for a limited time.
U
- UAT
- User acceptance testing: real users or the client check the product works for them before it goes live.
- UI
- User interface: what the product looks like on screen, including layouts, buttons and text.
- Unit test
- An automatic check that a small piece of code works correctly.
- Uptime
- How much of the time the product is working, often shown as a percentage like 99.9%.
- User story
- A short description of a feature from the user's point of view, such as: As a manager, I want a weekly summary.
- UX
- User experience: how easy and pleasant the product is to use from start to finish.
V
- Velocity
- How much work a team usually finishes in a sprint, used to plan the next ones.
- Version control
- A system that keeps every version of the code, so the team can see changes and undo mistakes. Git is the most common.
- Virtual machine
- A computer that runs as software inside a real computer, often rented in the cloud. Often shortened to VM.
W
- Webhook
- An automatic message one system sends to another when something happens, such as a new payment.
- WIP
- Work in progress: tasks that have started but are not finished. Too much WIP slows everything down.
- Wireframe
- A simple sketch of a screen's layout, made before the real design.
- Workaround
- A temporary way around a problem until the proper fix is ready.
X
- XML
- An older text format for storing and sending data, using tags. Still common in many systems.
- XSS
- Cross-site scripting: a security weakness where attackers sneak harmful code into a web page.
Y
- YAML
- A simple text format often used for settings and configuration files.
Z
- Zero-day
- A security weakness that attackers find before the software maker has a fix.
- Zero-downtime deploy
- Releasing a new version without the product going offline for users.
No terms match that. Paste it into Plainlee for a plain explanation.