None. I didn't say Notch can't ship a blockbuster game (and he can ship exactly one). I said he can't code worth a damn, and should not be idolized for his prowess.
Go and get developers at various companies. Not necessarily game companies, get someone that is payed a salary for "coding".
Give them some rough specifications (for example, a simple maze, simple lighting) and ask for Dart/WebGL to be used, and a timeframe of one month.
See what happens.
Based on my experience, 5% of developers may present something like it.
So, ok, Notch may produce spaguetti code, bugs, etc, but he's probably in that top 5% of developers. (Which is also debatable because someone may be good at, let's say, web development but couldn't code a game to save their lives)
Very few folks are idolizing him for his code correctness, but for his ability to prototype and ship.
I've worked with developers like Notch. I like them because they can create rapidly iterate apps/sites/ideas that normal people/users understand and critique. Then developers like me come in and polish the turds under the covers so we can ship and maintain.
He's not writing SQLite or Space Shuttle code, it's a amusement.
As amusing as your wordplay might be, the point being made is that Notch's coding skill should not be idolized or used as a reference for writing good code. A tangible consequence of his inability is that Minecraft takes up far more resources than it reasonably should and is very difficult to maintain, which has been echoed by several Mojang devs who have since picked it up.
I don't think they'll ever move away from Java. I don't think Minecraft requires something quite on the scale of a complete rewrite, but perhaps starting from scratch and pasting in bits and pieces of the original to save time would help things.
They are working on rewriting substantial portions of the codebase, but unfortunatley Notch didn't hand it off to the most skilled maintainers and they aren't making things much better.
I think this line of thought is a big problem in the development community.
There is a clear separation between code quality and shipping a successful products. Ideally you would want both if possible, but a product is not just code.
Ideally both, but if I had to sacrifice code quality for a successful product I would sort in exponential time if it meant getting a working product out the door faster. It doesn't need to be good, just good enough. (IMO obviously)
