Live data from Hacker News

From r60 to Go 1

gophersays.com

1–10 of 30 posts

Re: From r60 to Go 1

#2
Wow these are significant changes! The entire development process has been cleaned up significantly, and not having to use Makefiles is a significant improvement, as well as default error type. Go team is dedicated to developers and developer process and it will pay off.

Re: From r60 to Go 1

#3

Wow these are significant changes! The entire development process has been cleaned up significantly, and not having to use Makefiles is a significant improvement, as well as default error type. Go team is dedicated to developers and developer process and it will pay off.

Agreed, this sort of attention to tooling and the more menial parts of the development experience is sorely amiss among many other contenders, and definitely makes Go appear more interesting to me.

Re: From r60 to Go 1

#4
Looks like Go is cleaning up. Love the language, but I won't commit to writing heavy-duty code in it until it either has sensible union support or sensible generic support. The type system is just too weak right now.

Re: From r60 to Go 1

#5

Wow these are significant changes! The entire development process has been cleaned up significantly, and not having to use Makefiles is a significant improvement, as well as default error type. Go team is dedicated to developers and developer process and it will pay off.

go build reminds me of ghc --make, it's great.

That and Go reaching a 1.0 release may help me get it considered for some projects at work.

Re: From r60 to Go 1

#7
post #6

What's wrong with makefiles and why is new Go convention better?

All the information required to build a go package is in the source. why write and maintain Make files when they are unnecessary? The Go project is all about making programming simpler.

Re: From r60 to Go 1

#10
post #4

Looks like Go is cleaning up. Love the language, but I won't commit to writing heavy-duty code in it until it either has sensible union support or sensible generic support. The type system is just too weak right now.

I actually thought that to start with but when you get used to the frankly wierd type system, it actually works extremely well without generics (bear in mind I use C# which is heavily generic programming). When you approach the problems with a Go mindset, I don't think I've found a single case that needed generic types.

I'm not sure of the value of unions now or ever as it brings what is effectively a compiler decision into user space and allows very unsafe, architecture dependent code to be accidentally written.

Post reply on HN