My opinion goes like this: java is a verbose and hard-to-change language compared to ruby/python/js/etc so people think "I'm going to create this extra layer/abstraction just in case someone needs to change/use it for something else later" while dynamic language guys just solve the problem at hand and then improve it later because it's painless to do so.
Agree. Java coders seem to have a habit of over-solving the problem with useless layers of abstraction. I recall trying to create a silly video editor in Java, using a common vendor framework for media files, simply opening the file then extracting the Nth frame involved instantiating 12 layers of classes, like nested Russian dolls, every layer just as useless as it's parent class. What a waste of time it was to consider Java.
It's more a problem relating to the fact that many projects and their code bases are designed and planned to last for over a decade if not two or more. One simply can't just make drastic changes back and forth, because literally, peoples lives are on the line(let alone how much financial capital is invested...) with lots of written code. I'd grab Java, C++, heck, even C over Python or Ruby, let alone god damn JavaScript for a project which is going to ship after approximated 5 years of development and which has planned maintenance for at least 10 years afterwards. Considering how we miss deadlines, I'd assume such a project would be maintained after 20 or 25 years. Yes, someone would still work with the code base in 2037. Imagine that, and that's not even an unrealistic stretch. Welcome to enterprise software development! Do that with Python, Ruby or JavaScript and I'll give you a cake or two.
"Improve it later because it's painless to do so?" I don't think so. Quick and dirty approach and then fix it later is just the start of future issues, broken designs, infinite refactoring.
I feel java helps a bit more on design scalability. Yes, maybe somewhat overegineered but it certainly does not hurt to think on the future
