Hacker News Reader

Comment by throw10920 | Hacker News Reader

Interesting choice. I'd think that as the context is "As many of today’s (2023+) coders do not have a formal CS/SE education" and the goal is education, a more popular language like Javascript or Python (or, heck, even PHP[1] /s) would be used, instead.

I've taken a look at Go, and while it does seem pretty approachable, it's definitely not nearly as common as Python/JS, and it's always significantly harder for me to learn a new concept when the examples are also in a language I'm unfamiliar with. Maybe that's just me, though.

[1] https://survey.stackoverflow.co/2022/#most-popular-technolog...

View on HN

I think Go is a pretty good choice for the purpose. It balances high-level ease of use and learning curve with decent access to the system-y parts of coding that are so important to databases. What you learn to do in Go will translate reasonably well to a true systems language if the user wanted to take database engine design to the next level.

Languages like Python or Javascript are so far removed from the system-y side of programming that the way you would implement the concepts in those languages would not translate to the way you would actually build a "real" database which is I think the purpose of the book. I think the objective isn't to teach the abstract concepts but how those concepts are expressed in real systems.

I'm a data engineer and I only know Python. It appears Golang hits a sweetspot on many metrics such as performance, parallelism, ease of use etc and since 2016 there's been a lot of new data products and tools written in Golang. So it makes sense to me that the book would use a popular language for the domain.

Databases really do push runtimes in such a way that I think it makes sense to urge folks to use a system level language, or something close to it. In particular it'd be hard to cover concurrency (and parallelism) properly using vanilla CPython or JS, and I think that would impinge on the lessons learned.

That said, it'd be an interesting read on how to make a DB in pure Python.

you may be interested in this https://github.com/avinassh/py-caskdb

I don't normally think of Node and CPython as in the same performance bracket.

Regardless, you can do some crazy things in Node. See these notes about Node and MySQL:

https://github.com/tigerbeetledb/tigerbeetle/blob/main/docs/...

In this context I meant that the parallelism primitives are different than what you'll get in a systems level language, which might make teaching those parts unnecessarily awkward and probably the concepts less translatable to other runtimes.

For sure they've got different performance profiles.

Very impressive what your group was able to do with Node, and continuing on with Zig. It's got me interested in learning more.

I could be wrong, but my perception is that Go is so opinionated that you'll either write idiomatic Go or use another language. So, from that perspective there's some goodness in using Go as a learner's systems language.

It's opinionated, but it's not _that_ opinionated IMO. You can write awkward Java, shiny C, or whatever idiomatic Go is. Most people I worked with were from .net/Java worlds, and learned just about enough Go to be able to subject others to their coffee bean ideologies.

There's somethings the compiler will fail on like unused variable and the likes, but for the most part you need added static analysis and style checking -- some of which ships with the Go compiler.

Reply on news.ycombinator.com