Hey everyone! I've been a developer for nearly 9 years, mostly focusing on back-end work. Recently, I've started using AI tools like GPT and Claude to assist with my coding, especially since I began freelancing. While I've had some success fixing existing applications with them, I'm struggling with setting up new projects, particularly with frameworks like Spring Boot. I feel like I'm constantly needing to guide the AI, and it still makes a lot of mistakes. I've looked into the official Spring Boot documentation and have been trying to incorporate best practices, but I'm not sure if my approach is wrong or if these AI tools just aren't ready for complex setups yet. Any tips or experiences you all can share?
2 Answers
You might want to break down your project into smaller, manageable parts. I focus on my frontend first, working page by page since I’ve already defined my layout. For architecture, I recommend you nail down your business logic, choose your tech stack, and design your directory structure upfront. Also, create a README and initialize your Claude.md. Then, tackle the project in steps, asking Claude to help you along the way. Oh, and remember to add tests and use version control to keep everything in check before merging your work.
Hey! I found that it really helps to first have Claude structure your project using context schemas. For example, if you’re working on a repository, spend some time developing a structure file before you dive into coding. This idea of using context schemas can enhance how Claude understands your project. There's more on this approach that you might find useful: [Self-Tracing](https://github.com/recursivelabsai/Self-Tracing).
How is this different from just using the plan mode directly?
Very interesting! I'll definitely check that out, thanks!
Interesting perspective! I thought I was being methodical, but maybe I need to be more patient. I’ve been structuring it from the backend with RESTful endpoints but hit a lot of issues when I tried to test with Postman.