[urthen]

I wrote a quick Node Proof of Concept for a high-throughput data processing pipeline before the "real" team came in and rewrote it in Java.

Not only did the "real" team try and gaslight me into thinking my implementation didn't work (It did, they copied it wrong), but it took 3 engineers twice the time it took to copy my work into Java as it did for me to do the original development. The kicker: The Java process was so bloated it took nearly 2x the compute resources to get the same throughput.

TL;DR I challenge every time someone says "This is complex/high throughput, so let's use Java instead of Node"

[throw51319]

Yeah but don't you find digesting, creating a mental model, and then recreating someone else's work (esp in a diff language) to be more complicated than just writing something of your own from scratch? I kinda do.

Also, just curious, what things from a data processing pipeline standpoint does Node do well in? What does it not do well in?

Would Java ultimately be the most scalable?

[mrath]

Yes, like you said for PoC node is perfectly capable. When people say complex and high throughput then things like say Kafka or ERPs, banking software come into my mind. What are some of the biggest systems running on node? Sorry not challenging but curious to know. I have not heard of any popular big systems written in node.

[throw51319]

Yeah I work on a complex and scalable system for a bank using Java with Akka.

Could Node be as scalable in this regard?