I'm diving into a big programming project and I'm feeling a bit overwhelmed since I don't have much experience with large projects. I'm torn between two approaches. Should I: 1) Develop a simple base pipeline to get some initial tests running, knowing that it'll need a lot of modifications as the project grows? It seems easier to get started this way but it might end up being too simplistic. Or 2) Should I think ahead and design a more complex structure from the beginning, even if I don't need all the modules right away? This might help me establish a solid design but could be complicated since I don't have access to the server or real datasets yet. What's your advice?
5 Answers
If you know you'll need a server/database eventually, make that part of option 1. You can abstract the database functionalities so that during initial development, you just return hardcoded values to keep things moving along. It’s often a good balance to aim for a minimum viable product while keeping future needs in mind. Being too rigid can lead to changes you didn't anticipate later, and it's usually better to keep things flexible initially.
Avoid getting lost in complexity too soon. I use a site called withmarble.io for project management. It gives checkpoints to help keep things simple and organized, so I don't overthink everything.
Thx for the recommendation!
I like to start with something small, often just a simple "hello" world. If I'm unsure how to tackle a problem, I create a proof-of-concept. It's just for figuring things out, so no need to stress about tests yet. Once you've got a handle on it, then you can focus on writing solid code and tests. Test-driven development works for me, but learning through doing can be really effective initially.
I like that proof of concept idea! Once I find a solution, I can focus on writing it efficiently. Thank you :))
I actually think option 2 is better. Having a high-level design from the start prevents chaos later. You wouldn’t want to end up like building a bridge from two different ends, hoping they meet in the middle. Experienced developers often follow this route implicitly and it usually pays off in the long run.
I definitely recommend option 1. Start simple and avoid over-engineering your testing setup. There's this concept called "premature optimization" that you should definitely read up on. You want to get rolling without getting bogged down in details too early.
Nicee thank you!
Makes sense, thank youu so much! I've been thinking about creating a database with placeholders, even though I don't yet know the data format. I might just start with a JSON file for now and refine it later.