Hacker News new | ask | show | jobs
by rnewme 544 days ago
Of course, but there is a difference between programing using sdk/library/api and some glue, and actually engineering some self standing solution, and everything in between that. I belive blind programer wouldn't fare well in fast churn ci/cd every minute startup, but would be amazing in embedded space with multi year release cycles.
1 comments
scarface_74 544 days ago
How many large applications are made with no dependencies or frameworks?

In another life, I worked on a SaaS app around railroad car repair billing.

These are the industry requirements.

https://public.railinc.com/sites/default/files/documents/CRB...

And this doesn’t include addendums, regulation updates, industry norms, extra requirements by clients etc

No matter how well structured the codebase is, it will become large and had a lot of classes.

Not to mention the underlying .Net framework

Have you worked with either the iOS or Android SDK? Do you expect people to know that entire SDK and a sufficient large codebase?

Autocompleting IDEs are not a new concept. It’s been part of Visual Studio since the mid 1990s.

link
rnewme 543 days ago
I perhaps didn't express my point clear enough. If you're developing net framework, ios or android sdk itself, you're trying to be consistent and well structured, and you'll probably mostly work in one part of the code base. Then app dev will again use your structured expected layout as part of his structured system. The point wasn't to know the whole code base, but rather that it's made in a way that it makes sense how to use nth thing after you use first few.

I developed ios apps, and it's vastly different experience than android or flutter. And then that is not even comperable to JS and dozen build steps, standards, and libraries that are countless and all with different approaches to how they're structured and built...

link
scarface_74 543 days ago
And even if you make it consistent, using domain driven names, etc.

The method names are going to be different and not easily guessed. Why wouldn’t I use an IDE feature where I put a “.” at the end of the class instance and use autocomplete to tell me the methods? Why wouldn’t I use an IDE feature where when I rename a method, it automatically renames all of the references in an entire project?

Unless you are using reflection, this can be done in a provably correct way in a statically typed language?

link