For those just starting development

This article is aimed at people who are about to begin programming, mapping out the areas commonly encountered early in learning, on a fact basis. A guide for when there is no clear sense of where to start.

1. The big picture of development learning

Software development is layered work. Areas commonly met at the introductory level:

• Operating systems and terminals — the lowest layer that runs code and handles files. Mainly Windows · macOS · Linux.
• Editors / IDEs — the tools used to write code. VS Code · the JetBrains family · Vim / Neovim and others.
• Version control — Git is the de facto standard, paired with hosting such as GitHub · GitLab · Bitbucket.
• Programming languages — depends on the goal. Web frontend leans on JavaScript / TypeScript, data and scripting on Python, systems and performance on Go · Rust · C++, enterprise on Java · C#.
• Build / package tools — vary by language. The Node camp uses npm / pnpm / yarn, Python uses pip / uv / poetry, Java uses Maven / Gradle.

2. A common learning order

There is no absolute rule for the order, but the following flow is often recommended. Introductory courses such as Harvard CS50 follow a similar structure:

1. Terminal basics — cd · ls / dir · creating / removing files · exiting the shell. The most basic interface for talking directly to the operating system.
2. Pick up one editor — VS Code is a reasonable starting point for beginners. Free, with rich extensions.
3. The basic syntax of one language — variables · conditions · loops · functions · data structures. Once one language sits comfortably, the next tends to come faster.
4. Git basics — init · add · commit · push · pull · branch · merge. Only sticks once typed by hand enough times.
5. Small projects — a calculator · a todo list · a simple web page · a text-based game, something with a finished result. Beyond following tutorials, a step of self-designing is needed.
6. The habit of reading official docs — blog posts are quick but often outdated or inaccurate. Official docs are the most reliable.

It is fine to jump to the next stage when stuck and come back later. Learning is not linear.

3. Resources to reach for when stuck

• Official docs — the primary source for languages and libraries. Hard at first, but the fastest path once familiar.
• MDN Web Docs — the de facto reference for web standards (HTML · CSS · JavaScript · Web APIs). Run by Mozilla.
• Stack Overflow — searching by the concrete error message often surfaces a similar problem and its fix. Look at the accepted answer, the vote count, and the date together.
• GitHub Issues · Discussions — for bugs and questions about a library itself, that repository's Issues are most accurate.
• Search engines — wrapping the concrete error string in double quotes cuts down noise.
• AI assistants — useful for quick explanations, less so when fact-checking is hard. Re-confirm the decisive parts with official docs.

4. Differences by environment

Even with the same code, paths · line endings · shell commands differ across operating systems. When introductory material assumes macOS / Linux, Windows users often need extra translation. The environment/ bundle in this collection covers those differences.

5. Common misconceptions

"You must master one language before moving on" — any language goes infinitely deep. Once basic syntax and one or two small projects are done, moving to other areas (web · DB · tools) helps maintain motivation.

"Strong math is required" — depends on the area. General web development needs only arithmetic. Graphics · machine learning · cryptography on the other hand need linear algebra · probability · discrete math.

"You need a powerful laptop" — basic learning runs on most entry-level machines. Specs start to matter only in heavy areas (machine learning · game engines).

"When stuck, search is the answer" — search is fast, but wrong answers arrive fast too. One more glance at the official docs saves time over the medium and long term.

6. Small signals of progress

• The habit of wrapping error messages in double quotes for searches sets in.
• Reading and following the "Installation" section of a README becomes possible.
• The first move on a new command is checking its help (--help · man · Get-Help).
• Personal code gets split into git commits.
• The second time the same problem appears, it is solved faster than the first.

When these signals accumulate, the introductory stage is mostly done.

Closing thoughts

Learning is not linear but a chain of jumps and returns. One language + one small project + the habit of reading official docs + Git basics — these four form the foundation of the introductory stage. When stuck, official docs are the faster route over the medium and long term, more than search. When the same problem is solved quickly the second time around, the introductory stage is mostly done.

Next

• environment/01-windows-vs-mac
• environment/09-first-terminal-day
• learning/04-learning-roadmap

We refer to The Missing Semester of Your CS Education (MIT) · Harvard CS50 · MDN Web Docs · Stack Overflow Help Center · GitHub Docs Git Handbook.