I only watch the go part, and I'll say that in 3 years working with I might have had at most 3 times nil pointer crashes in prod, in which took about 30 min between getting fixed and deployed.

There are linter which helps prevent most of if not all crashes (just keep in mind to run linting and compile the binary it would still be ages faster than anything rust I have ever compiled). His argument is weak, and not simple.

I'll give that type system in golang is too simplistic sometimes, and a more complex could help to express better some use cases.

Still go for a person coming from a interpreted language is a solid choice by being MUCH MUCH simpler.

▲

pythops a day ago | parent | next [-]

I agree on the easiness of Go, but I don't think people should always choose the easy solution. The easiness of Go come at a cost.

	▲	gabrielgio a day ago \| parent [-]
		Agree, people should understand why they are picking a language. If it is to learn new language that is fast, compiles to a binary (also fast) and has a nice onboard experience for a person coming from a interpreted language (which is the topic of thread) go fits nicely. Rust is not always the answer just because it has "no nil exception". The correctness of Rust comes at a cost as well.

▲

super_flanker 20 hours ago | parent | prev [-]

I agree with simplicity and ease of learning Go.

> 3 years working with I might have had at most 3 times nil pointer crashes in prod

I've been running a rust app for my personal trading app and a small service at a very large FAANG company for more than 3 years, and guess what I'm yet to see a nil crash.