What Deserves Its Own Package?

Not every class or object deserves a namespace. Ask yourself:
• Can it exist independently, or is it tightly coupled (like UserProfile inside User)?
• Does it provide a meaningful interface, or is it just an internal detail?
• Does it evolve on its own, or always with its parent?

The answer guides whether you create a subpackage. Micro-level autonomy matters.

👇

Macro Level Matters Too

At the macro level, namespaces still help: separating layers like .domain or .infrastructure enforces isolation and clarifies architectural boundaries.

Thus, we can see it as a positive thing rather than a meaningless technical label — these boundaries are, for the most part, self-sufficient.

👇

Isolation Principle

Even small pieces follow the isolation principle. If it’s not independently meaningful, it shouldn’t get a package. Example: UserProfile is tightly coupled to User — it’s not a standalone concept, so it doesn’t deserve its own namespace.

👇

Singular vs Plural

Packages represent concepts, not collections. Use singular:
• ✅ user → all code related to the User concept
• ❌ users → implies a collection, which can be misleading

👇

Conclusion
• Say no to .common, .core, .shared, .base
• Say no to .model, .dto, .entity
• Yes to meaningful, concept-first packages
• Avoid dumping unrelated code into generic folders
• Only split code if it’s independently meaningful and exposes a clear API
• Focus on the concept first, layer second

👇

Don’t Fear Many Files!

More files aren’t a problem if each placed correctly in its meaningful place. A well-structured codebase is easier to navigate than one full of generic “drawer” packages.

Check out my full article for deeper insights and examples.

I’ve just updated my website — y9vad9.com.

It looks noticeably better now and supports three languages. Articles are now available in English, German, and Ukrainian. There’s still room for improvement, so feel free to let me know if you spot anything.

I’m also planning to publish at least two new articles before 2026. The topics are quite interesting, so stay tuned.

Sometimes a function signature doesn't tell you what it wants. You read the implementation, figure it out, and blame the missing docs. But are docs the real problem? Or, maybe, the code says String when it means something much more specific.

New post on semantic typing in Kotlin — when it helps, when it gets in the way, and the reasoning behind design decisions that usually happen without much thought. Roughly, going from "this is a string" to "this is a concept".

Enjoy! https://dev.to/y9vad9/semantic-typing-we-ignore-i0d

And not to forget — wishing y’all great holidays! Merry Christmas 🎄

By the way, I also today updated one of the first articles I ever wrote — the one about Extension-Oriented Design, which I also recommend checking out: https://dev.to/y9vad9/extension-oriented-design-3d41

Hey again, fellas!

As promised, I've published a new article — about contracts, contract violations, and how to think about failures in code.

When exceptions are the right signal, when they’re harmful, and why "safe" abstractions can sometimes be more dangerous than throwing. We'll also look at patterns established in the libraries we use every day — when they actually work, and when they quietly don't.

Hope you will find it useful — https://dev.to/y9vad9/failures-we-dont-model-correctly-2hbo.

Say goodbye to your mouse 👋

I’ve just shipped a new command palette to the website. It’s fully keyboard-driven and works everywhere — from the Home page to any note — making navigation fast and frictionless. Just press '/' on any page!

The palette surfaces:
• Mentions and links referenced inside a note, so you can jump around without touching the mouse
• A clearer view of cross-references between notes
• Tags — a newly introduced part of notes. Type tag:something to instantly find notes from a specific category

More details here on palette you can find here: https://y9vad9.com/en/notes/site-features


On top of that, I’ve refreshed the notes pages:
every article now has a preview image, visible both in the list and on its dedicated page.

That sounded unreasonably epic for something you'll probably visit once in a lifetime, but I still love it.

Hopefully soon I'll finally get my hands on publishing some of my notes (not quite articles, so the only real place for them is my website). Then this change might actually make sense. 😁

New small note:
• UI shouldn't think about validation

P.S. The website underwent significant changes to adopt a 'digital garden' mental model. So, don’t be surprised — I will post an update about it once it's fully transformed.