The practical glossary
Understand the word.
Connect the idea.
Plain explanations, concrete examples and the mistakes to avoid. Find the missing piece, then get back to building.
56 terms · A to Z
- Building a project
Absolute path
An absolute file path identifies a location from a filesystem root rather than from your current folder. Syntax varies by operating system: macOS and Linux paths start with /; a Windows example is C:\Projects\reading-list\README.md.
See an example → - Working with AI
AI chat
A conversational interface where you send messages and receive generated replies. It can help explain, plan or draft work. Whether it can also use tools depends on the application and permissions available in that session.
See an example → - How the web works
API
A defined way for software to interact with other software. It describes available operations and the expected inputs and results. A web API often uses HTTP requests, but APIs also exist inside libraries and operating systems.
See an example → - Identity and security
API key
An API key identifies an application or service when it calls an API. Its permissions depend on the provider. Some keys are designed for public clients; privileged secret keys belong only in trusted server code.
See an example → - Running a product
Application email
Messages an application sends outside its own interface, such as sign-in links, confirmations or receipts. Sending involves a mail service and delivery process; request acceptance, delivery and a person's action are separate observations.
See an example → - Identity and security
Auth
A common shorthand for an application's identity and access machinery. Authentication establishes who is making a request; authorization decides whether that person may perform the requested action. Sessions carry identity between requests.
See an example → - Identity and security
Authentication
Authentication verifies who a caller claims to be. A sign-in system checks evidence such as a password or a verified sign-in link, then establishes an identity. Knowing that identity does not automatically grant access to every record or action.
See an example → - Identity and security
Authorization
Authorization decides whether a caller may perform a particular action on a particular resource. It uses rules such as ownership, role or purchased access. The trusted server or database enforces those rules; hiding a button only changes the interface.
See an example → - How the web works
Backend
The application logic running in a trusted server environment rather than a visitor's browser. It processes requests, applies business and access rules, and works with data or outside services on the application's behalf.
See an example → - Running a product
Background worker
Here, a program that receives background work and performs it outside the visitor's immediate request. It needs a defined completion signal and failure behavior. The same word can mean other things, such as a browser worker or provider runtime.
See an example → - Running a product
Backups
Preserved copies or recovery information used to restore state after loss or damage. Source code, database records and uploaded files may need different backup methods. A recovery plan also specifies what to restore and how to verify it.
See an example → - Shipping changes
Branch
A branch is a named line of development in Git. Its reference advances as you make commits. Branches let you work on a change separately before combining it with another line of development; they do not isolate external services or live databases.
See an example → - Shipping changes
Build
A build prepares source files for delivery, often compiling code and producing files or server functions a host can run. The exact steps depend on the project. A plain static site may be served directly without a separate build step.
See an example → - Data and storage
Cache
A retained copy of data or a result used to avoid repeating slower work. A cache needs rules for its key, freshness and replacement. Its contents can differ from the underlying source after that source changes.
See an example → - Working with AI
Coding agent
An AI workflow that uses available tools to work toward a coding task, such as reading files, proposing edits or running checks. Its authority comes from the tools and permissions granted to it, not from the model's name.
See an example → - Shipping changes
Commit
A commit is a recorded snapshot in Git history, with a message and links to earlier commits. It records the content selected for that commit. Making a commit does not automatically upload it to a remote or deploy it for users.
See an example → - Shipping changes
Content delivery network
A distributed network that delivers content from locations closer to visitors. It often caches eligible responses, reducing repeated trips to the original server. What can be cached and for how long depends on the configured rules.
See an example → - Shipping changes
Continuous integration
A workflow that automatically runs defined checks when code changes are integrated or proposed. It can build the project, analyze code and run tests, giving you evidence about a particular source version.
See an example → - Data and storage
Database
An organized collection of data that an application can query and update. A relational database organizes records into tables, connects them through keys and can enforce constraints that protect valid relationships and values.
See an example → - Building a project
Dependency
A dependency is software your project relies on. In a JavaScript project, package.json lists required packages and version ranges. Some packages support the running app; others help develop, test or build it. Their compatibility matters alongside your own code.
See an example → - Shipping changes
Deployment
A deployment puts a particular version of your application into an environment where it can run or be served. A preview deployment is for checking a candidate; production serves your users. Deploying code and changing a database are separate operations that must remain compatible.
See an example → - How the web works
Domain name
A human-readable name in the internet's naming system, such as example.com. Registration controls the name; DNS records describe destinations, and hosting serves the website. A domain is only one part of a complete URL.
See an example → - How the web works
Domain Name System
A distributed lookup system that connects domain names with records, including addresses and mail-related information. Nameservers answer for a domain's configured records; different record types serve different purposes.
See an example → - Building a project
Editor and IDE
An editor lets you work on source files. An integrated development environment, or IDE, combines editing with tools such as debugging and building. AI assistance can be added to either; it is a separate capability.
See an example → - Building a project
Environment variable
Environment variables supply named configuration to a program. A .env file is one way to load them. They can hold ordinary settings or secrets; the variable name alone does not control who receives the value.
See an example → - Data and storage
File storage
A place to keep file contents such as photos, recordings or documents. Object storage identifies each stored object by a key within a bucket and can attach metadata and access rules to it.
See an example → - How the web works
Frontend
The part of an application people interact with: its content, controls and visible states. On the web, browser-side code handles interactions, while some page content can be prepared on a server before it reaches the browser.
See an example → - Shipping changes
Git
Git records versions of a project's files and connects those versions into a history. You can inspect changes, work along separate branches and combine work deliberately. Git runs locally; a hosting service such as GitHub is a separate place to share repositories.
See an example → - Building a project
GitHub
A platform that hosts Git repositories and provides collaboration tools, including issues and pull requests. Git tracks your project's history; GitHub holds a remote copy and helps people review and coordinate changes.
See an example → - Running a product
Google Search Console
A Google tool for website owners to inspect how Google discovers, indexes and presents their pages in Search. It reports issues and search performance, helping you investigate visibility rather than guessing from your own search results.
See an example → - Shipping changes
Hosting
The service or infrastructure that makes a website or application reachable over a network. It may serve prepared files, run server code or combine both. Your project must fit the host's supported runtime and deployment model.
See an example → - How the web works
HTTP
HTTP is the shared set of rules clients and servers use to exchange requests and responses. It carries pages, images and API data. HTTPS protects that exchange with encryption; it does not decide whether a user is allowed to perform an action.
See an example → - Running a product
Logs
Records of events emitted by software while it runs. Useful logs include safe context such as time, operation and source version, helping you connect an observed failure to what the application actually did.
See an example → - Shipping changes
Merge
A merge incorporates work from another line of Git history into the current one. Git can combine compatible changes automatically. When changes conflict, you must decide the intended result; making the merge finish does not prove the resulting application works.
See an example → - Data and storage
Migration
A database migration is a recorded change to the database's structure or data. Migration files let a team review and repeat changes in a known order. Applying one changes the database; committing the file alone does not apply it.
See an example → - Running a product
Monitoring
Watching selected signals about a running system to detect failures or changes that matter. Availability checks, errors, measurements and alerts can reveal different problems. Each alert needs a clear condition, owner and useful response.
See an example → - Running a product
Payments
The workflow for collecting money through a payment provider and handling the resulting state. Checkout, payment confirmation, refunds and application benefits are related stages that need explicit rules and reliable evidence.
See an example → - Running a product
Product analytics
Collecting and interpreting defined observations about product use, such as page visits or completed actions. Events and funnels help answer product questions when their meanings, collection boundaries and data quality are clear.
See an example → - How the web works
Proxy
An intermediary that receives network requests and forwards, redirects or modifies them according to configured rules. A reverse proxy sits in front of services; some applications also run request logic near the network edge.
See an example → - Shipping changes
Pull request
A pull request proposes merging changes from one branch into another on a repository hosting service. It gathers the change description, discussion, review and automated checks. Opening one presents the work for a decision; it does not itself merge or release the change.
See an example → - Running a product
Queue
A system that holds messages or tasks for consumers to process separately from the request that created them. Delivery, ordering, retention and retry guarantees depend on the queue and its configuration.
See an example → - Identity and security
Rate limiting
A rule that restricts how many operations a caller or group may perform within a defined period. It helps control overload, abuse or spending. The counting key and limits should fit the operation being protected.
See an example → - Building a project
Relative path
A relative file path describes a location from a chosen starting folder, usually your shell's working directory. It does not start at the filesystem root. The same path text can point to different places when the starting folder changes.
See an example → - Building a project
Repository
A Git repository stores a project's recorded history and references such as branches and tags. Your working files are the version you currently edit. A remote repository stores shared history elsewhere; it is not automatically the running application.
See an example → - How the web works
Request
A request is a message asking a server to do something. An HTTP request identifies a destination and method, carries headers, and may include a body. It describes what the caller wants; it is not proof that the caller has permission.
See an example → - How the web works
Response
A response is the server's reply to a request. An HTTP response includes a status code and headers, and may include a body. Read the status and returned data together to understand what the server actually reported.
See an example → - Running a product
Rollback
A deployment rollback sends traffic back to an earlier application version. It can help recover from a faulty release, but it does not automatically reverse database changes, payments or other external actions. The earlier version must still work with the current data and services.
See an example → - Identity and security
Row Level Security
Row Level Security, or RLS, lets a database enforce rules about which rows a caller may read or change. Policies depend on the caller's identity and role. Privileged roles can bypass these rules, so RLS does not replace every server permission check.
See an example → - Running a product
Scheduled job
A task invoked according to a configured schedule rather than a visitor's click. Schedulers commonly use cron expressions. Timing, retries and possible duplicate invocations depend on the scheduler, so the operation needs suitable recovery behavior.
See an example → - Identity and security
Secrets
Sensitive values that grant authority or reveal protected information, such as private service keys or signing material. They need controlled access, appropriate storage and a lifecycle that includes rotation and revocation.
See an example → - Data and storage
SQL
SQL is a language for working with relational databases. You use it to define tables, request matching rows and make changes. A query can filter, combine or order records so your application receives the data it needs.
See an example → - Shipping changes
Staging environment
An environment used to verify changes before they serve production users. It aims to represent relevant production behavior while using deliberately chosen configuration, data and service destinations for safe testing.
See an example → - Building a project
Terminal
A terminal is the text interface where you enter commands and see their output. A shell interprets those commands and starts programs. The terminal window and the shell inside it have different jobs, even when an editor presents them together.
See an example → - Building a project
Users
The people using your product to accomplish something. They may visit without an account, use different devices or rely on assistive technology. Design and evaluate the product around their tasks, not only around the developer's screen.
See an example → - How the web works
Webhook
A webhook is an HTTP notification sent to a configured URL when an event happens in another system. Your receiver verifies the sender and handles the event. Delivery can be repeated or delayed, so receiving a notification is not the same as finishing the work.
See an example → - Building a project
Working directory
The working directory is the folder a running program uses as its current location. In a shell, it is the starting point for ordinary relative file paths. Different terminal sessions can have different working directories, even within the same editor.
See an example →