Worked, and Yes. But thats due to legacy more than anything else, and because the builder tools team built a pretty sizeable ecosystem around dependencies/builds with Java, with a lot of templates out of the box.

>tell us more about how using java mean we dont know computer science. still want to understand how language choice a “computer science” problem. teach us.

Early on at Google, there was an engineering decision to use “Python where we can, C++ where we must.” There is a reason for that.

In terms of performance, a VM language is really good when you are dealing with multiple architectures. Java was a good choice back in the day to support application development across multiple environments.

As such Java was the natural choice for Android for quick application development. But lets look at how that evolved over time

- Most performance heavy applications are written heavily with NDK. Google provided this specifically because the jvm layer doesn't suit things like games very well.

- Java isn't even the official SDK language, Kotlin is preferred, which essentially "fixes" a lot of Java things. BTW there was guidance at Amazon to use Kotlin over Java going forward because it results in a better maintainable codebase.

- Google is heavily investing into Fuschia, with Flutter/Dart, as the next gen mobile/home device OS. In general a lot of web app application style apps are written with React Native, which is structured very similarly to Flutter/Dart (as the code can be compiled to a JS web app as well)

So it seems like Google, the company responsible for putting a JVM on majority of the worlds smartphones, doesn't even believe it it anymore. Python or Golang is all you really need these days for any backend - you develop faster, code runs just as fast (especially with Python 3.11, or previously with PyPy), taking account that most of the time the latency driver is going to be network requests, and you have a much better tooling with those language.

As for vulnerable software you are correct though , vulnerabilities can happen in other languages. Golang seems to have quite a few. Python had one back in 2007 with tarfiles. Meanwhile, your favorite Java now has Spring4Shell circulating around, and I hopefully don't have to explain how in the case of Log4j, a developer willingly writing the network request/execute code as part of logging something is much, much worse than finding some bug that can make a program crash in very specific conditions.

And this is because of the core issue of Java of having so few core libraries compared to Go and Python - so instead you have random projects with questionable people working on them, and then these projects get all hodgepodged together into the modern day Java best practices. So you have things like big dependency injection frameworks that eat startup time, or things like Lombok that essentially hack the AST. There is no central authority controlling what Java is, so you get these Log4Shell, Spring4Shell and others.

But lets ignore all that, and focus on the fact that Java still runs fast, for whatever weird reason you need speed. So what exactly does Java give you over Golang or Python? OOP semantics? Strong typing?

Lets look at how that works in practice. Every professional Java codebase will have unit tests. Everything that you can possibly ever catch compile time you absolutely can catch with unit tests, and there will never be an argument against writing less unit tests for anything because "compiling hypothetically catches all the errors for you". Now take your standard java project, where you have a build system like Gradle. You run your compile process, which is groovy being compiled to JVM that gets executed to run additional commands that do things like annotation processing, after which all of your code gets compiled into a jar file, then ran, where the dependency injection frameworks have to load, and then your unit tests get ran. At Amazon, every time you wanted to either run unit tests or start up the service, it took at least 10 seconds for everything to run before your code gets executed. In comparison, launching Python interpreter to run your additional unit test that you wrote is quite fast. Look at AWS Lambda start times between Java and Python if you don't believe me.

The only thing strong typing does is that it makes it harder for, lets say "entry level" devs to shoot themselves in the foot, since they are essentially using an API. But Java does provide ways to shoot yourself in the foot, which is why code reviews are a thing. And if you have code reviews, then you can collaborative drive a certain structure even in dynamically typed languages like Python towards minimizing mistakes, even more so because development happens faster so you have an easier time putting in everything that needs to be put in.

So in the end, if you didn't realize all of this, then you don't fully understand computer science. If you do realize this, and still think Java is ok to use today, then you either are delusional, or are privy to some great wisdom that nobody else knows.