I think such complaints are the predictable backlash that's correlated with the final stage of its hype-cycle. Tech hipsters always need a hot new thing to tout so that they can pretend to know what they're talking about.
Now that we have a new iteration of hyped languages in Swift, TypeScript, and Elixir, Scala, Go, and node.js are on the hipsterdom downtrend (which often also means they're on the mature user uptrend). That prior generation of languages are now at least semi-widely known, so they no longer serve the purposes of the talentless hipsters, who base their self-worth on the esoteric nature of their preferences and need something that most people don't yet know much about to hide their ineptitude.
I think Python is a great example of a community that built a real niche for itself and legitimately earned the respect of mature, senior engineers by repeatedly proving and improving its utility and stability for many years. Giving Go the opportunity to develop this same type of sustainable, consistent, healthily growing ecosystem without a flood of 0.1x programmers and other hipsters is really the best thing for it. I'm optimistic for Go in the long term.
Look behind hype cycles and often you'll find $company marketing department.
It's not a criticism, it's literally how the Go language creators themselves describe it. See the Rob Pike quote at http://nomad.so/2015/03/why-gos-design-is-a-disservice-to-in...
I'm responding to criticisms in this and other HN threads about Go, so yes, I'm talking about a criticism.
> it's literally how the Go language creators themselves describe it. See the Rob Pike quote
Unless you think Rob Pike is saying that Google is hiring mediocre programmers (hint: he's not), that statement is intended as a criticism of academics ("researchers"). He's saying that the same things that excite researchers about programming languages don't lend themselves to good software engineering in industry. (And he's right).
But regardless of what the creators of the language may have intended, I'm saying that the criticism regarding experience with functional languages is flat-out wrong. I've written more code in various functional languages than the overwhelming majority of people who read this site - both on and off the clock - and I'm testifying that yes, even for someone who has deep experience with functional language of both the Lisp and the SML families, Go still has a lot to offer that very few others do.
Actually that's exactly what he's saying - although "inexperienced" is probably a better word than mediocre.
Golang is the modern blub language. It's really great for large companies and big code bases used by many people who want to get quickly up to speed. A decent, even new, programmer can get up to speed and be very productive in Golang in a matter of weeks. I can go and read the Kubernetes code base, which is huge, and know exactly what's going on pretty quickly, despite the complexities in that codebase like code generation and massive concurrency.
A language like Rust on the other hand would take months before someone can be seriously productive in it. That has a very real cost. It's sacrificing short term productivity for the long term benefits.
I think for a big company like Google, with so much staff, many inexperienced, and relatively high turnover (lots of people leaving and going constantly), a language like Golang is perfect.
At a small hedge fund, where you have maybe a dozen or two lifers, investing in a more strategic language like Haskell, Rust, OCaml, etc. starts to become viable.
Because if Golang maintains its current momentum, someday soon the developing world will start pumping out legions of mediocre Golang coders, just as they did with Java.
Personally I don't think there's such a thing as a 'mediocre' programmer, I think there are programmers who have studied and trained in concepts, techniques and idioms and those who haven't.
It seems pretty clear from that quote that Google is hiring kids fresh out of college, and instead of training and educating them they're throwing them at an intentionally simplified language to get them to churn out code quickly.
Also I find it hard to believe that type-safety guarantees don't lead to good software engineering in industry, when the biggest industrial users of Python, PHP, JavaScript and Clojure--the dynamic language titans of the tech industry--have all recently developed and popularised type systems for them.
And in fact, I also don't see how you can derive a criticism of academia out of those quotes when they mention 'researchers' only in passing, as a point of comparison to talk about their coders' skill level. Too much mental gymnastics for me.
Which is really just talking in circles, because as I've now said in three consecutive comments, I have quite extensive experience with functional programming, and so I'm rejecting the argument that Go is simply for people who don't have that experience.
> And in fact, I also don't see how you can derive a criticism of academia out of those quotes when they mention 'researchers' only in passing, as a point of comparison to talk about their coders' skill level.
It's probably because, in five years of writing the language, I've had the opportunity to hear Pike (and others) talk about the topic. So when I'm responding, I'm drawing on that body of knowledge to inform what I'm saying, rather than limiting myself to a single quotation cited on a blog post written by someone who (by his own admission) is very new to the language.
You may reject it, but the fact remains that that was explicitly one of their design goals when creating Go.
> ... a single quotation cited on a blog post written by someone who (by his own admission) is very new to the language.
A quotation is a quotation; either Rob Pike said those words or he didn't. Who cited the quotation doesn't make a difference.
If that was the case, it was a very dubious thought. Before Go was launched, the languages used in Google were Java, Python and C++. None of them look academic, none of them seem to require another language that is specifically designed to "unexcite" newcomers. Neither is functional, really.
Sounds like someone's ready for Haskell.
I don't know how else to describe people who don't resent being second-class citizens under the library designers, who don't insist on being able to create their own abstractions and use them with the operators that slices or maps support. The language has extensible interfaces but doesn't use them for things like iteration and equality.
One suggestion: describe them as "people who have a different opinion than I do", rather than automatically concluding that anyone who doesn't feel the same way as you do about something must be "mediocre". There are popular languages that I can't stand writing, and languages that I think are badly designed, but I would never be tempted to say that those languages are "for mediocre programmers" just because they don't have the attributes that I look for.
As a polyglot, I can definitively say that every language in widespread use is "missing" something in its design that, at first glance, would be glaringly obvious to someone who's used to a different language. If anything, the sign of a good programmer (as opposed to a mediocre one) is the ability to conceptualize why a language might be successful despite (or even because of) this apparent "flaw".
"for...range."
"How do I iterate over my own type?"
"The hard way."
"What interface can I implement to make for...range work?"
"You can't. Our types are important enough to support properly, yours aren't."
As a programmer I expect more than that, and I don't respect people who meekly agree that creating types is properly left to their betters.
Some "different opinions" are just different opinions, other "different opinions" do indeed point to mediocre programmers.
I'm not saying that is the case with Go: just that what you wrote is orthogonal. One can have a different opinion that is NOT to be respected, but dismissed. Not everything is equally valid.
I've been writing Go full-time as my primary language for nearly five years. Before that, my languages of choice were Lisp[0] and Python, with R as a very distant third choice[1]. I have always been a polyglot and enjoyed experimenting with any new language I could try out - you'd be hard pressed to name a non-esoteric language that I haven't written more than a "Hello, World" in at some point.
I still write other languages and appreciate them for what they are, but Go is what I reach for when I need to crank something out. The language gets out of my way and doesn't distract me - and as an added bonus, I can be reasonably confident that I can quickly refactor the code 6 months later and still have it run[2]. Go is not a scripting languages, but I sometimes even write small applications in Go that I would have previously written as shell scripts, just because it's easier and less thorny than remembering to avoid all of the pitfalls of shell scripting.
You might say that this is because I've been writing the language for so long - and yes, I'm definitely more comfortable with it today than I was in 2012. But everything I just said was still true even when I was only a few weeks into learning the language. It was like putting on glasses for the first time in my life - yes, I could "see" before, but somehow everything was just a little bit crisper, and I felt incomprehensibly more powerful and capable when writing Go compared to Lisp, Python, Scala, Java, Perl... despite having far more years of experience in those other languages.
If some people don't want to use Go, that's fine with me. But I strongly reject the criticisms that Go is meant for "mediocre" programmers, for programmers without experience in functional languages, or somehow an inferior language because it lacks feature X or Y from language Z.
[0] Common Lisp (via SBCL) or Racket, depending on the project and collaborators
[1] It's a language with a lot of warts, but even Python couldn't touch R for completeness in statistical libraries.
[2] Seriously, I have never found a language that made refactoring as easy - or, dare I say fun? - as Go.