|
|
|
|
|
|
by sriram_malhar
584 days ago
|
|
|
I fully agree with you about the solidity of the VM. The others I am not so sure about. > Bottomless resources of developers with Java experience. With Java experience, but what fraction have a systems outlook? What fraction have an experience with other languages to ensure that the code they write is simple and understandable and direct? My own experience is that too many come out addled by Enterprise Java idioms, and when you actually write some code in Erlang or Go you realize systems aren't as complicated as they have been made out to be. > Managing very large codebases ... I wonder if this is self-fulfilling. My theory is that these codebases are huge because their designs are enterprisey. The primary drivers of complexity are indirection: factories, dependency injection, microservices, these are all part of the same malaise. |
|
|
It depends what you mean by systems outlook but JVM based code is pretty common (to the point I’d say ubiquitous) in large distributed systems.
In open source it’s much the same. Many of the large Apache projects are in JVM languages, for example.
> The primary drivers of complexity are indirection: factories, dependency injection, microservices, these are all part of the same malaise
The indirection in Java does drive me crazy. But dependency injection is a problem to solve in every language and libraries that can do code generation at compile time like Dagger2 make this predictable, debuggable, and fairly easy to reason about on the JVM.
Microservices are, in my opinion, more of a business organization solution than one tied to any specific language. If you haven’t read Steve Yegge’s blog post about Amazon vs Google I think it’s good reading on why/when SoA is a good idea.