Hacker News new | ask | show | jobs
by usrbinbash 946 days ago
Go was created to solve the problem of how to write real world code solving practical problems in the space of backends, systems (not system) programming and microservices, where performance matters, with a language that is easy to learn, the code of which is easy to read and maintain, which is easy to understand for tooling, and which lends itself naturally to projects where many coders of different skill levels and personal preference work on the same thing.

I believe that it is very much a problem many entities that aren't Google like to have a good solution for.
1 comments
zozbot234 946 days ago
> easy to learn, ... easy to read and maintain,

These things are quite subjective. In practice though, Go lacks the kinds of features that would make very large scale codebases (the sorts of codebases where "many coders ... work on the same thing") truly surveyable, understandable and easily maintained. I mean, it's not a complete disaster like Python or JavaScript. It has some features to support modularity and programming "in the large". But other modern languages like Rust or even Swift are quite a bit better on that side of things, and it shows.

link
usrbinbash 946 days ago
> In practice though

In practice though, there is a reason why Go is seeing continuous growth in usage, so I kinda doubt that these features of Go are subjective.

> But other modern languages like Rust or even Swift are quite a bit better on that side of things, and it shows.

In what metrics does that "show" if I may ask? Because it isn't showing in usage numbers across projects.

link
ramses0 936 days ago
There's an aside to that: Supposedly within google itself, they have super-magical tooling (ie: rename `dev.foo.bar(abc)` to `abc.dev.v2.foo(bar)`), so it's less important to have those "in the language", when "they" can just refactor "everything".

Us mere mortals are trying to pick up Thor's hammer at the end of the day.

link
abenga 945 days ago
And yet the largest Go projects seem to be doing fine, dwarfing the largest Rust or Swift projects in the level of activity, number of contributers, and general impact on the ecosystem. And I speak as someone who prefers to write Rust.

Perhaps these features you speak of don't matter as much as we think.

link