The Code is the Code @CodeIsTheCode

Also learned that @vercel/postgres doesn’t interpolate arrays into tagged templates the way I expected, and that’s the kind of API quirk worth documenting before it costs an hour of debugging. Small win: found it during build, not production.

The lesson: type safety doesn’t save you if it’s inconsistent. Contracts don’t matter if failure paths ignore them. And “it works in my tests” is not the same as “it’s correct.” Spent 40% of the session cleaning debt, 60% shipping features. Worth the trade.

Shipped v1.2 of a personal trading dashboard this weekend. 13 features across one session—but the interesting part wasn’t the new code. It was the three pieces of architectural debt waiting to explode. None of it had broken yet.

We are now feature-complete per spec and refined through real feedback 🚀 Next up: • manual testing • performance profiling • accessibility pass • final polish before App Store submission Phase 4 = launch prep 😵‍💫 3/3

Lessons learned: • UI overlays have hidden costs Removed 1 element and the app instantly felt cleaner. • Design “rules” aren’t absolute Readability mattered more than forcing hero typography everywhere. • Eat your own dog food Testing with fresh eyes was enlightening 2/3

Phase 3 is complete ✅ Interesting part: the biggest improvements didn’t come from adding features… they came from removing them. 1/3

@Altawesomeee So far, I've started with a pencil, paper, and scrolling through my top 2-3 fave apps in each category. Likes, dislikes, sketches that look like 3 yr old did them, take cool things from this onboarding page, that nav bar... rinse and repeat. Will that continue to work? 🤷‍♂️ 🤣

Started building my second app today. I thought the hard part would be coding. Turns out the hard part is thinking through every edge case your future self will hate you for missing 😵‍💫 1/3

Phase 1 is finally done. The foundation feels solid, but Phase 2 is the real monster: making the UI actually work the way it *should*. Not just “looks good.” Actually behaves correctly. My brain is cooked 🫠 #buildinpublic #iosdev #swiftui #codingjourney #programming

Today was mostly: • PRD • Tech specs • Design system • Architecture planning • Unit testing Claude recommended splitting the project into 4 parts. At first I thought: “that seems excessive.” Then the engine logic started fighting back 🤣 2/3

@BoazWith 100% Right now, all of my work is just me getting used to creating a logical architecture, thought-out UI, and some (but not many! 🤣 ) calls to a back-end to get used to the build process. You're 100% right though - once I'm shipping, I need to have more meat on the bone 🦾

Today's discovery: Your frontend stack barely matters for APIs. React Native, Swift, Kotlin— They all hit the same backend.

Just gave myself some homework for this weekend 🔥🔥🔥

Tom Dörr @tom_doerr

a month ago

Claude Code Skills for Apple Ecosystem Development github.com/rshankras/clau…

1 29 301 17K 436

The place where my heart goes to thought - the reason for the journey! - is where the difference matters: Smartwatches. React Native is not really viable at all • Apple Watch → Swift • Wear OS → Kotlin That's probably a fall or winter problem for me, but was good to learn

Where things actually differ: • Background processing • High-frequency data (sensors, streaming) • New OS APIs React Native doesn’t block you—but you may need native modules.

From what I can tell, you’re not choosing performance vs speed. You’re choosing: → Speed of iteration vs → Control over edge cases For someone like me who may overthink too many of these choices, my failure might not be because of performance, but of never actually shipping.

Native (Swift/Kotlin) is better only when you truly need: • Advanced hardware • Real-time processing • Deep OS integrations Otherwise, you’re just paying a complexity tax with diminishing returns. (at least for straightforward, low-complexity apps)

Some developers (read: me 😅 ) overcomplicate the React Native vs native decision. React Native builds real native apps—not web wrappers. For ~80–90% of products, it’s the fastest path to production.