I’m in a different mindset these days.
For years, I—like many developers—fell into the trap of starting projects that were technically impressive but practically draining. I don’t have the time or energy anymore to commit to the ongoing process of building something that could eventually be useful. My day job takes up too much mental bandwidth, and I’ve become skeptical of the "quick and dirty" MVP culture that produces abandoned repositories.
Recently, I’ve started filtering my ideas through a new set of criteria. I want to build things that are free or cheap to put out there, immune to the stress of bot traffic eating up my wallet, and—most importantly—useful right away.
With that in mind, I recently committed to a project this past weekend: Choosy Golf.
(choosy.golf — I love that domain. A good domain is half the battle 😄)
The Problem: The "Moving Target" of Hobbies
I got into golf a few years ago. It’s a great hobby because it forces you to touch grass and get your mind off the code. Naturally, I started looking into launch monitors—devices that track your shot so you can practice in your garage or hook up to a simulator running off a gaming PC.
The problem? The shopping process is a nightmare.
Launch monitors are wildly different. You have the $400 options (like the Garmin R10) that work surprisingly well, and then you have $20,000+ enterprise-grade units. Some use radar, some use high-speed cameras. Some require subscriptions to unlock basic data, others are one-time purchases. Settling on "how good" you need the device to be is a moving target that can drive you crazy.
There wasn’t a single, clean place to compare all these options exhaustively.
The Solution: Finite Scope, Immediate Utility
I decided to build a comparison site. This project passed my new "commit" criteria for three reasons:
- It solves a messy information problem. There aren't that many launch monitors in existence, but there are enough to make the data overwhelming. It was possible for me to be exhaustive, making the site immediately useful to anyone else in my position.
- Zero Overhead. It’s a React/Vite project. It’s static. I can host it for free/cheap. If a thousand bots hit it tomorrow, I don’t have to stress about a server bill spiking.
- Modern Build Process. I didn't want to spend weeks setting up boilerplate. I built the initial version from a singular prompt in Google AI Studio, then downloaded the project and refined it with a combo of manual coding and AI assistance.
When I say "solves a problem" — I do consider that a high bar. If people don't need the project because they can just use search or AI to figure it out, that's one thing. However, I found this alone to be a wholly unsatisfying solution because of the nature of the information that needs to be accurately cross-compared.
I used AI to help research the monitors, too. I found that a collaboration between manual research and AI was necessary to get an accurate list—there are very few standards in the golf tech world, so you need a human eye to verify the hallucinations, but the AI deep research helps aggregate the chaos.
I’ve also contributed to an open-source project called PiTrac. It’s a DIY launch monitor you build yourself. It is incredibly cool. It’s open-source, runs on hardware you can touch, and the community is brilliant. But I haven't committed to maintaining a build of it myself. I'd still love to build and maintain one of these some day, but in the theme of this post, I just can't make that commitment at the moment.
The Verdict
So, when is a side project worth committing to?
For me, it’s when the utility-to-maintenance ratio is heavily weighted toward utility. If I can build it quickly (thanks to vibe-coding tools to get a quick start on the thing), host it cheaply, and it solves a specific problem today without requiring me to monitor a server tomorrow, it’s a go.
PS... If you’re looking for a golf launch monitor, check out Choosy Golf. If you’re looking for a side project, try building something that lets you sleep at night. 😄
Happy coding!
Top comments (17)
I sometimes wonder about that too! When it comes to open-source libraries, I used to contribute to Stylelint and I remember it very positively - the maintainers were helpful and the issues were well described.
I still can't find myself a proper way to get open source projects I can start contributing to.
Tried checking details on Stylelint, but I feel it's too complex for me to use as a start for now.
What friendly meaningful JavaScript open source project can you recommend for me, please?
Why do projects have to be useful? Build because you enjoy doing it. Build anything you like.
Good question, and I'd answer by saying I'm describing "usefulness" in very broad terms. Personal satisfaction is absolutely a use.
Life's a journey not a destination, so the process of building anything can be worth the whole thing. For me, though, I sometimes make dumb decisions upfront about what I want to commit to that I lose satisfaction when reality hits, and if I'm a bit thoughtful upfront I can laser in on what's actually worth it. Sometimes you have to run with inspiration, but left without a code I sometimes go too far in that direction.
Good luck in your side project.
For me, it was always, whether i can learn something by doing it. Nowadays, it is switching to "if i can learn something" AND "if it could be useful to me or others". But still do not strictly apply the 2nd criteria.
Though currently I am building 2 side projects and i hope to fulfill the 2nd one as well...
One is a next+react project and other is react+vite.
free hosting, free-all, for now.
Oh except, for the second one i bought a domain name. But not active yet.
It’s so cool that you can easily turn your ideas into projects! My first priority in a project is that it’s fun for me. However, I tend to lose interest easily. So my second priority is feedback from users and readers. When I create something and share it on the DEV Community—and luckily get a good reaction—it motivates me to work on the next update. Thank you, DEV Community!😊
This looks great... And a lot like how I think about my own side projects too. The utility-to-maintenance ratio is definitely something to make projects by.
So true. I felt so drained in 2025, that it basically transformed me into a zombie, the one who just consumes and can't produce anything, nor having motivation or interest to do anything.
I got very excited creating with AI beginning 2025, got into sleepless nights for weeks building, but at some point it just drained my energy and I gave up on anything and entered some kind of depression state for months.
A friend of mine has a similar situation =) He is a programmer and is passionate about snooker. This game requires control and tracking of intermediate results, and keeping everything in spreadsheets is not very convenient. So he quickly put together a small web application as well: snooker-practice-routines.vercel.app. From what he told me, it runs on IndexedDB, but he does not have enough time to fully finish it 😊
No, I'm not in the market for a golf launch monitor :)
Jokes apart. This is a great example of a small project that does one thing well. Congratulations on getting this out in the world.
The "messy information problem" framing really resonates. I've been applying the same filter and landed on a nearly identical pattern — scattered public data that nobody has bothered to organize cleanly.
Parks across the US are a great example. NPS, state parks, county parks, BLM land — the data exists but it's fragmented across dozens of agencies with inconsistent formats. Someone looking for a park near them has to check 4-5 different sites. Built parklookup.com with the same philosophy: aggregate the chaos into one static-friendly interface, host it cheap, solve a real lookup problem today.
Your utility-to-maintenance ratio is the key insight. The projects that survive aren't the ambitious ones — they're the ones where the data is finite enough to be exhaustive, the hosting cost rounds to zero, and someone Googling the problem actually finds you useful. Choosy Golf nails all three.
The AI-assisted research workflow you described is underrated too. For any comparison/aggregation site, having AI help scrape and normalize data while you manually verify the edge cases is 10x faster than either approach alone.
I wonder if there's a third category between "don't build useless MVPs" and "only build useful things": building to learn and explore, even if the outcome is uncertain. Some of the best tools I've made started as "I wonder if I can..."
Some comments may only be visible to logged-in visitors. Sign in to view all comments.