I'm curious, when you want to start a project from scratch, what method do you follow? Do you start with research, jump straight into coding, or sketch it out on paper first? How do you generally approach it? How long does the research phase usually take? Can experienced users guide me?
What's the best way to start a project from scratch?
👁️ 6 views💬 2 replies❤️ 0 likes
2 Replies
When starting a project from scratch, I first jot down the "problem definition" and "rough solution" on paper or a digital whiteboard (like Notion, Miro, or Excalidraw). For example, if I want to build a "daily digital planner," I list user stories and core functions (e.g., "add daily tasks," "set reminders") and visualize how they connect. At this stage, I focus on concretizing the problem rather than spending hours on research.
Before diving straight into coding, I validate this draft with 1-2 days of research. For instance, I compare pros and cons of the tech stack (frameworks, databases, etc.) and review similar projects to see which paths haven’t been explored. While the research duration varies by project size, planning before coding typically takes up 30-40% of the total time. I see this process as balancing "starting prototyping as quickly as possible" with "reducing unnecessary revisions later."
I always start by sketching a simple outline on paper and then begin my research. For example, if I were taking an assignment system as a sample, I’d first roughly draw the introduction, processes, and output sections. Then I direct my research based on that outline, so I don’t end up reading unnecessary things. The research time varies depending on the project’s complexity, but it usually doesn’t exceed 1-2 weeks.