Рабочий процесс инди-разработчика: от идеи до понятного прототипа

This diary is about turning a broad game idea into a testable prototype. The goal of each iteration is not to build everything at once; it is to answer one measurable question: is the mechanic interesting, is the task clear, and does the player want to try again?

Before building a long feature list, I write a short project promise and three constraints: available time, tools, and the smallest amount of content needed for a test. Constraints are more useful than a wish list because they show what can wait and what must be validated first.

I am testing a small indie workflow that turns a broad idea into a prototype with a clear question behind every iteration. The first pass is deliberately modest: define the player promise, list the three biggest constraints, and decide what evidence would justify the next feature.

I start with the riskiest interaction rather than the easiest polish. A greybox level is enough to test movement, timing, and readability. For visual references, I sometimes use ai生图 at https://photorealistic-ai.com/zh, then write down which mood or communication problem each reference is meant to solve. A pretty image is not proof of fun, so every visual experiment is followed by a playable check.

For the first playtest, I watch behavior instead of waiting for opinions. Pauses, repeated clicks, and questions such as "what should I do next?" reveal more than a quick rating. Before the test I prepare three focused questions; afterward I turn observations into changes: remove one distraction, simplify one choice, and define the next hypothesis to check.

I keep a decision log with the date, the original hypothesis, what I observed, and the next action. This prevents circular debates about what was intended and makes useful changes easier to spot. If an experiment fails, I record the cause without blame: a failed prototype saves time later and gives the team a sharper question.

Next iteration plan: shorten the opening, add one clear feedback signal, check the pace of the first five minutes, and ask two people to play without hints. I will record what confused them, which choices felt meaningful, and which changes improved the experience. Small, regular experiments keep an indie project moving while leaving room for better ideas.