I'm confident you've also ran into large messy Rails apps too :)

Rom has some other advantages, like not needing a 1-to-1 mapping between tables/columns to models/attributes. For example, you could have a `users` table that maps to Profile and Account structs. Profile could include a `bio` whereas Account version could have the email, ensuring data doesn't leak beyond what's needed. They're two different use-cases for the same data, so should be modeled separately. This also lets you work with DB's where you don't control the schema (owned by other teams or legacy DB's): you can map the data into the structure you need for your app, instead of carrying around irrelevant attributes.

I tried out Rails Engines on a couple projects, with such high hopes and ran into issue after issue. Sure it's theoretically possible to build whatever you want with Rails, but in practice it's infeasible. Some people have experimented with arbitrarily nesting Hanami slices too and had success. It's not something we're focusing efforts on because it's rather niche and get inherently complicated but it's possible.

Modularity is such an important part of large software projects and Rails doesn't give you any tools for it (and Ruby doesn't help either). Packwerk was an attempt to constraint Rails, with limited success: https://shopify.engineering/a-packwerk-retrospective.

Hanami doesn't have any specific Inertia or Svelte specific integrations, but our view layer (and assets layer) is very flexible so it's certainly possible. It's worth investigating. If you try it out, let us know what you learn on our forum https://discourse.hanakai.org/

[1]: https://github.com/realworld-apps/realworld

We re-did our homepage recently, and we should make these things clearer. They're legitimate concerns and questions. We agree you don't have to write bad code in Rails, and we don't pretend that we can prevent bad code in Hanami. Instead we give application builders the tools to architect their applications in the way they desire. Really what we're doing is building an *option* for building Ruby apps that may speak better to some people. If you're completely happy with Rails then there's no reason to change. Still, you may learn something that'll help you build Rails apps by seeing how we do things. Additionally, it's completely modular: you can pull parts of Hanami into a Rails app. The most popular one people pull into Rails apps is dry-validation [1]

Some concrete differences:

- We have small, simple API's. Rails core classes have hundreds of methods for convenience, not to mention dozens of core extensions (a.k.a. monkey patches) to base Ruby classes. We prefer small, narrowly focused classes over large ones with many responsibilities.

- We have a dependency injection framework dry-system [2] included by default. This lets you write `include Deps["repos.article_repo"] to get an automatically instantiated (and now memoized!) `Repos::ArticleRepo.new` within your class. This makes dependencies trivial to stub out, and it also declares your dependencies in a single place, instead of littering constants throughout your code.

- We have sub-applications called Slices. Similar to Rails Engines but well... good, since they're a first-class concept that we encourage. Importantly, they're independently deployable. You can export and import components across slices, too. We can't tell you how to slice up your application, but we give you the tools to decide for yourself.

[1]: https://hanakai.org/learn/dry/dry-validation

[2]: https://hanakai.org/learn/dry/dry-system