The Go Programming Language

by snewe • 17 years ago

377 218 17 years ago

Chrome. Closure. Go.

It says great things about Google that these all clearly started as projects that they assumed nobody would ever have to Google. (Although http://www.google.com/chrome is now #1 for [chrome]).

jkincaid• 17 years ago

Is there a reason why so many languages are nearly impossible to search for? (C, C++, Go, etc.) I mean, surely someone must have figured out that this is not the best naming trend quite a while ago.

83457• 17 years ago

http://www.google.com/search?hl=en&q=%22[[INSERT_PROGRAM...

mdemare6• 17 years ago

Our kind is so dominant on the net that our interests eclipse the original concepts. Try googling Java, Python, Lisp or Ruby - not much about Indonesia, snakes, speech impediments or gems to be found.

joeld423• 17 years ago

Really? http://tinyurl.com/y8zjorj

frognibble• 17 years ago

Rich Hickey picked the name Clojure partly because the name is easy to search for.

sans-serif• 17 years ago

Those are nothing compared to d and r (relatively unknown) and my favorite: io.

mnemonik• 17 years ago

Have you ever tried to find anything about Io (http://iolanguage.com) through google? Impossible. Any projects I think up in the future are going to have to be named from this angle: how easy will it be to google for?

msluyter• 17 years ago

Searching "go programming language" (without the quotes in the search) seems to work for me.

andreyf2• 17 years ago

Hm, it could be a self-challenge - make web search good enough to know what I'm talking about when I look for Go?

I'd expect: the game, then the language. Certainly not go.com

whatusername• 17 years ago

Since I saw this the other day:

http://duckduckgo.com/?q=go&v=

ynniv• 17 years ago

Maybe its a branding exercise in disguise. You can't search for ["Chrome", "Closure", "Go"], but for the first two you quickly get what you're looking for if you associate the company name: "Google %s".

snprbob86• 17 years ago

It's legally the safest and easiest route. If your company name is already trademarked, you don't need to worry about someone tacking a common word on the end and stealing it. This is why so many Microsoft products have such generic names -- and all of them include at least one prior trademark. Google is at least trying to keep a flat namespace instead of Microsoft's nested hierarchical mess. If nothing else, it would help prevent duplication of efforts because of potential naming conflicts.

rayvega• 17 years ago

It seems another programming language named Go! already exists:

http://news.ycombinator.com/item?id=935545

prabodh• 17 years ago

I am just wondering how same link got posted twice ... http://news.ycombinator.com/item?id=934140 ....Both point to http://golang.org ..

mdemare6• 17 years ago

Looks interesting! I feel that an improved C is what's mostly lacking in the programming language landscape nowadays.

So now Google's got its own OS (two actually), browser, and now its own programming language. Anything missing yet? Their own editor / IDE?

andreyf2• 17 years ago

Their own editor/IDE?

I wouldn't imagine this as something that they'd release.

ecuzzillo• 17 years ago

Oh, you never know. They release a lot of the not-business-specific infrastructure they improve, which an editor undoubtedly is.

I'd argue their secret sauce, and the parts they wouldn't want to release, are - the insane scaling stuff - all the intelligent things they do with the insane amount of data they have, and the insane rate at which they can process it

viraptor2• 17 years ago

It's like several people got the same idea... http://ooc-lang.org/ is also an improved C really.

SwellJoe• 17 years ago

Bram Moolenaar (vim) is a Googler.

mbrubeck2• 17 years ago

...and has his own experimental C language replacement: http://www.zimbu.org/

blasdel9• 17 years ago

They actually maintain at least four browser codebases: Chrome, Android's Webkit, Picasa (uses Gecko), and their server-side renderer.

They have at least three product OSes: ChromeOS, Android, and The Google Search Appliances.

This is at least their second new programming language project: there's also Rob Pike's Sawzall. They have numerous in-house implementations of existing programming languages: V8, Closure, GWT, Dalvik, Unladen Swallow, plus all the extant ones they maintain forks of.

Slashed2• 17 years ago

May I ask you? What do you mean by improved C language? Safe pointers? I think this is the biggest advantage of C.. Of course, it can(and occasionally it will) become a trouble at some point, but you won't get that much flexibility and portability in any other language. Not trying to say that this is a bad idea to create a new language or anything, but I just don't see how this language will make my life better..

P.S. No offense, but it's amazing how some people are such Google fans.

frankus• 17 years ago

I can't find any sort of reference to a Google sponsorship. Am I missing something?

agl5• 17 years ago

We (the Go team) all work at Google :) See also http://www.youtube.com/watch?v=rKnDgT73v8s

messel• 17 years ago

Thanks Adam.

akamaka2• 17 years ago

I'd have to disagree with the title "Python meets C++". There's very few features that have been brought over from either of those languages.

It looks like C with some very carefully chosen additions: memory safety, concurrency and maps.

mbrubeck2• 17 years ago

I agree. From the FAQ:

"Go is mostly in the C family (basic syntax), with significant input from the Pascal/Modula/Oberon family (declarations, packages), plus some ideas from languages inspired by Tony Hoare's CSP, such as Newsqueak and Limbo (concurrency). However, it is a new language across the board."

Two of the designers were Rob Pike and Ken Thompson, both instrumental developers of Unix/C and Plan9/Inferno. Thompson created the B programming language that C was based on.

dhess• 17 years ago

Don't forget Russ Cox, who I believe stuck with Plan 9 longer than either Rob or Ken.

I'm excited to see CSP in Go. I've been advocating its use for concurrency since I learned about Plan 9's channels, but most people paid no attention.

ypavan• 17 years ago

From the Tech Talk PDF, it is mentioned that Robert Griesemer, Ken Thompson, and Rob Pike started the project in late 2007. Ian Lance Taylor and Russ Cox joined in 2008.

And the goal, as they mention, seems to have the efficiency of a statically-typed compiled language with the ease of programming of a dynamic language but with -

Safety: type-safe and memory-safe.

Good support for concurrency and communication.

Efficient, latency-free garbage collection.

High-speed compilation.

They are aiming at a lightweight type system with no implicit conversions and no more 0x80ULL.

A good starting point is their FAQ page and also the tech talk at youtube: http://www.youtube.com/watch?v=rKnDgT73v8s

dejv• 17 years ago

Syntax looks like Pascal meets C++

chickamade• 17 years ago

I'd say that name-based interface is indeed Pythonic.

> Any type that has the named methods [...] is said to implement the interface.

For instance, in the tutorial, the cat function takes a reader interface. fmt.Print and Println a type will output whatever its String() "method" output. This is what Python does, e.g. "print x" is eqv. to "print x.__str__()".

silentbicycle2• 17 years ago

That's not a Python thing - it was present in Smalltalk, for example.

rsc• 17 years ago

A much more interesting addition is the interface type and values.

rbanffy• 17 years ago

As any pythonista would say: "from __future__ import braces"

japherwocky• 17 years ago

garbage collection also.. which is maybe the most "pythonic" thing they've worked in. look at all those semicolons and curly braces!

tptacek2• 17 years ago

I am not so interested in Go as an implementation platform, so much as I am in the idea of a bridge for web programmers to real C.

mdemare6• 17 years ago

I like C, maybe it is Good Enough, but don't you think C has some real flaws that could be fixed in a newer language?

tptacek2• 17 years ago

I think there's ~30 years of C code that isn't going away and that is going to remain the bottom layer of most programming stacks, and it's good for devs to be able to drop into it.

benreesman• 17 years ago

I was ready to be dismissive until I read who was behind it. This is part of the Bell Labs legacy. Kickass.

viraptor2• 17 years ago

I also suspected bell labs the moment I saw the page... not sure where that idea came from. Maybe from this? ;) http://plan9.bell-labs.com/plan9/img/plan9bunnyblack.jpg http://golang.org/doc/logo-153x55.png

[edit: creepy thought... Go logo == the Glenda bunny being thrown at something?]

Semiapies2• 17 years ago

Perhaps flying? :)

amichail3• 17 years ago

How come who was behind it is important but who says what on HN is not?

One could have extra information right beside each submitter/commenter's name (e.g., ... founder, Google employee, PhD in ..., etc.).

tspiteri• 17 years ago

You can say that who is behind it is important because you do not need to add extra information besides their names, their names stand on their own.

DougBTX• 17 years ago

There are some usernames I will recognise, and their comments do mater more to me. Also, if someone does say something interesting, you can often learn more about them from their profile page should you wish to.*

And, yes, it would amuse me for a couple of days if everyone with/working on a PhD were encouraged to append their thesis titles after their usernames.

* though yes, for people I don't recognise who are not posting particularly interesting comments, it does not matter who posted them. Such as this comment I assume, anyone could write it.

fragmede• 17 years ago

There's also follow through and prolonged effort which matters on a project vs. points for discussion. On HN, the idea is the point each comment makes should be considered on its own merits. On the opposite end of the spectrum; the project should stand on its own merits. However, those merits are likely to be derived from the choices of the person(s) working on the project, so the person(s) involved in the case of a not-yet-finished project is relevant.

omouse2• 17 years ago

It's tainted because of that. I became dismissive as soon as I saw who was behind it...

gloob2• 17 years ago

Intriguing. Mind if I inquire about your thoughts on transistors, another Bell Labs creation?

apgwoz• 17 years ago

As soon as I got to the channels portion of the tutorial, I knew it was Rob Pike's baby. He did a tech talk a while ago about parallel languages, which was very interesting http://video.google.com/videoplay?docid=810232012617965344&#...

EDIT: after watching some of the tech talk, it's evident that while Rob Pike is involved, it's not entirely his baby... He is however involved, and there are definitely ideas from newsqueak in the language.

c00p3r2• 17 years ago

You probably know what happened to NeXT legacy.. =)

tumult4• 17 years ago

I use C when I have a program with tight timing and/or memory requirements. I see that this is a GC language, so am I right in assuming that it's not suitable for soft or hard realtime systems?

agl5• 17 years ago

GC pauses are current unbounded, so I wouldn't use it for that sort of thing, no.

tumult4• 17 years ago

Right on, good to know, thanks :)

kaib2• 17 years ago

Most of Google infrastructure is soft realtime so GC delays have definitely been a concern. The memory footprint for a simple hello.go on arm using 5g is currently ~70k with the optimizer not yet enabled. I'm planning to do some work on a AT91SAM7 with 256kb of flash and it should fit OK. You might be able to squeeze it into a smaller space if you drop some of the runtime.

tumult4• 17 years ago

Assuming the OS threads/preemptive multitasking are playing nice (i.e. letting me devour timeslices at my leisure), how well do you think the GC would work when I need to land events in a 15 microsecond window at 40hz on a normal desktop PC? I'd be crunching a decent amount of math and discarding/creating objects between events, but total memory utilization would not be going up or down.

Sorry if that's kind of specific, it's some of the limitations in one of the current problems I'm working on :)

davidw• 17 years ago

Erlang is used for 'soft real time' applications, and it does GC.

nearestneighbor• 17 years ago

AT FIRST GLANCE:

Annoyances: braces; variable declarations more verbose than in C

Flaws: no discriminated unions; no generics or parametric polymorphism

Doubts: no exceptions; no macros (as far as I can tell)

    Edit: I take back the verbosity claim. := does inference
    Edit2: interfaces replace generics, but are 
        type-checked at runtime, apparently:
        http://golang.org/doc/go_for_cpp_programmers.html#Interfaces

agl5• 17 years ago

No macros indeed. Unions types will probably happen sooner rather than later.

jrockway• 17 years ago

Yeah. Languages with no control flow other than "call" and "return" were ok for the 80s... but we know better now. Just add continuations and you get functions, coroutines, exceptions, loop exits ("last", "next", etc.), and so on for free.

I am not sure why I'd use this language with no libraries when I could use a better-designed language with no libraries instead.

telemachos2• 17 years ago

It ain't CPAN yet, but I think "no libraries" is an over-statement: http://golang.org/pkg/

btilly3• 17 years ago

There are other control flow options available. For example they offer the following simple example of what they call goroutines:

    ch := make(chan int);
    go sum(hugeArray, ch);
    // ... do something else for a while
    result := <-ch;  // wait for, and retrieve, result

You can actually set up any complex mess you want of functions talking to other functions through channels. This gives you coroutines, threading, etc. Looking through the spec, it looks like they have anonymous functions and closures as well.

While I agree that they need to add exceptions, they do have some interesting things in there already.

gchpaco2• 17 years ago

No macros is pretty annoying, but note the package go/ast and go/parser; they include in the standard library a parser for the language, which should make writing tools operating on Go significantly easier.

blinks• 17 years ago

On Edit2: interfaces don't replace generics, they're more like a strong duck-typing system (if that makes any sense).

Generics are being worked on (according to the tech talk), but they haven't determined the best way to do them yet.

mleonhard• 17 years ago

Exceptions are definitely needed.

mdemare6• 17 years ago

Garbage collection, type inference, built-in maps and arrays, immutable strings, concurrency support. Very nice!

Best of all, no header files!

thaumaturgy• 17 years ago

As a random aside: it's possible to build C++ projects without header files, and get your source files to automatically work out their own dependencies and all that good stuff.

I did it with a horrible C++ preprocessor hack. It was fun.

dchest2• 17 years ago

And for C, see makeheaders http://www.hwaci.com/sw/mkhdr/

milkshakes• 17 years ago

http://golang.org/doc/go_faq.html#Concurrent_programming "Do not communicate by sharing memory. Instead, share memory by communicating."

Luyt2• 17 years ago

Ars Technica on Go

http://arstechnica.com/open-source/news/2009/11/go-new-open-...

Luyt2• 17 years ago

It seems like Ars Technica also have their own shortened links...

http://arst.ch/9uw

mdemare6• 17 years ago

Nice article, but calling Go not innovative because it has a C-like syntax is a bit naive.

wendroid2• 17 years ago

Journalists journalist, coders code.

dschobel• 17 years ago

Despite the large amount of enthusiasm for language design, modern mainstream programming languages don't fall far from the C tree. The best that Microsoft, Sun, and Apple have to offer are just variations on that theme, with the addition of predictable object models and conveniences like garbage collection.

Hey, give MS some credit for F# here! And it's got out of the box support in Visual Studio 2010.

jhawk28• 17 years ago

Doubt it will be successful, no beards: http://www.youtube.com/watch?v=rKnDgT73v8s

agl5• 17 years ago

Ken, iant and I have beards thank you :) Rob, gri, Russ and Kai are sadly lacking.

nailer19• 17 years ago

Any chance you could please ditch the brace requirement?

You're already marking up the code for humans using indentation, don't repeat yourself by having the compiler use a different parsing mechanism (and make developers mark up seperately for their colleagues and the compiler).

blasdel9• 17 years ago

"Go has its own model of process/threads/light-weight processes/coroutines, so to avoid notational confusion we call concurrently executing computations in Go goroutines."

I like this not just because it's CSP reborn -- new languages should be introduced with their own delightful disambiguating nomenclature.

nailer19• 17 years ago

I like it because it rhymes.

Freebytes2• 17 years ago

I am not interested in really learning another programming language, and from the looks of it, the productivity claim is questionable. Not only can many of the examples they show be done in fewer lines of code in other languages, but you also do not need to re-learn the languages you already know. Many programming languages offer many of the features they are mentioning as useful so why learn yet another programming language unless it truly has something useful to offer that does not already exist. As one person stated, it has braces. I like braces for the purpose of organization, but I dislike variable declarations because it only serves to clutter the code and sometimes decreases productivity. The programming language needs to become more productivity oriented, and I do not think Go accomplishes this goal. Developers want to be creative with their code, but they would probably rather be more creative with their projects instead.

jacquesm10• 17 years ago

This is hacker news, learning new programming languages and learning about them or thinking about them is part of what hacking is all about (at least for me it is).

If you think that programming languages are a 'done deal' then try to ignore the last 15 years or so and see what you're left with.

Sure, we're not much closer to the holy grail of programming in absolute terms, but new languages (erlang, clojure, ruby, python to name a few) have definitely incrementally changed the perspective we have of programming, and have opened up more fruitful ways of building certain classes of applications.

Freebytes2• 17 years ago

When I stated "I am not interested in really learning another programming language", I should have worded it more clearly. "I am not interested in learning this particular programming language because it does not seem to be a step above existing languages. It seems outdated before it has even been released. Simply because something has big names behind it does not mean the product is automatically superior to others. I intend to learn other languages, of course, but I do not intend to waste my time.

joeld423• 17 years ago

wow... first impression is this is my dream language... Can't wait to start messing with this...goroutines yay...I could be wrong but this looks really nice.

vaksel• 17 years ago

So how long before we start seeing craigslist ads that ask for 5 years of Go experience?

jurjenh• 17 years ago

not before the end of the world (2012)

stcredzero• 17 years ago

Perhaps time travelers exist, and this is just them accidentally blowing their cover?

jjs• 17 years ago

2 weeks, tops.

Krechet• 17 years ago

What the hell??? Didn't they come out with programming language last week already or some such? I don't see the point exactly. I mean I can learn it, but then what? It's not going to be installed anywhere where I can do work, it's not going to have any third party libraries, it's not going be be supported by any decent editor for a while (I use vi, throw your rocks).

Google: what is the point? First it's browser plugins, then browsers, then two OS's, now programming languages? How about a nice new processor architecture a la PowerPC? It's better and faster you know. Maybe they'll also invent new people to be their user base. I can imagine that intro video: "Welcome to GPeople - the new Google interaction experience. When you interact with a GPerson you don't need to talk. They already know what you are thinking and will insert relevant advertisement into their conversation with you. The advertisement will be unobtrusive and will target your innermost fears..."

dchest2• 17 years ago

They already support your editor: http://code.google.com/p/go/source/browse/?r=release#hg/misc... :-)

rikthevik• 17 years ago

Google has a ton of really smart people using their 20% time on projects they find interesting. I think they've shown great restraint in what they release publicly. I'm sure there's a ton of very cool new things being worked upon in the company.

omnigoat3• 17 years ago

My word, a whole bunch of what I'm reading is exactly what I have/had intended to implement in my toy language, prism. Bit-specified integer (and real) types, increment/decrement operators as statements (prefix only!), hell, even using "var <X> ..." for variable declarations. The guiding principles are divergent though. And the bit about deducted type relations (http://golang.org/doc/go_lang_faq.html#inheritance) leaves me with mixed feelings.

I'm actually happy! It means my stupid little language has some validity :P

For those who wish to deride it: http://code.google.com/p/prism/

omnigoat3• 17 years ago

Well, in the interests of actually furthering discussion:

I feel as if string and map simply could have been part of the standard library. My solution to the this for prism was to have a section of the standard library (which doesn't exist yet) marked as "integral", to effectively have an implicit "import string, map from std.__integral__" at the top of every translation unit.

The absence of overloading operators seems to contradict their claims of improving the visual "flow" and feel of programming. There are cases where savvy operator overloading works wonders: vectors/matrices/quaternions, strings, complex numbers, etc. To be able to write my_string + my_other_string is far more elegant than strcat(my_string, my_otherstring). Consider what happens when you have four of five strings you'd like to concatenate!

The choice of CSP reflects Google's mentality: Their problems lend themselves very well to CSP. Other domains (in my opinion) work better with a STM model. But that's clearly a design choice, and I think it's fine!

agl5• 17 years ago

Operator overloading has good points and bad points: we're erring on the side of simplicity for now.

However, you can write "abc" + "def" in Go and it does what you expect.

oconnor0• 17 years ago

I'm not a huge fan of "+" for string concatenation as "+" is, traditionally, commutative; I like Lua's "..". That being said I like "+" more than "strcat".

wheels• 17 years ago

After reading "The Case for D" a little bit back, what these leaves me wondering is not, "Why this over C or C++?" but "How does this compare to D?"

After a quick glance, D piqued my interest more. D seems to fall a bit nearer to the C++ side of the fence than Java, but also boasts C++-level performance and relative safety.

http://www.ddj.com/hpc-high-performance-computing/217801225

Neither of them will replace C or C++ in the near-term simply because those languages are so entrenched, but if I were looking for a language just to play around in for potential future use from these families, at this point it'd be D.

michaelneale2• 17 years ago

Looking good. I always thought something like this would be LLVM based, this is surprising.

9oliYQjP• 17 years ago

They said in the FAQ that they "felt" LLVM was "too large" and "too slow". I would have liked to see this statement backed up with data.

nikils2• 17 years ago

link to Ken Thompson's previous c-compiler paper http://doc.cat-v.org/plan_9/4th_edition/papers/compiler search for 'Performance'

michaelneale2• 17 years ago

Yes I saw that (and took their word for it). It is interesting that they seem to deeply care about compile time speed (I guess given Go is a fresh start in so many ways, people thought it would be nice to have super fast compiles for quick turn around and testing).

vicaya• 17 years ago

The home page touts "safety" as a feature the language but all it takes to crash the program is concurrent access to the builtin map: "uncontrolled map access can crash the program", according the faq.

Interesting so far besides some syntax ugliness and lack of macros, but not yet good enough for many system programming projects, especially given the state of the garbage collector. GC is the hardest thing to implement large-scale, concurrent, memory intensive applications (e.g, database servers). Java spent 18 years to perfect its GC to be the current state of the art, which still sucks for these applications.

I hope they can come up with a concurrent garbage collector that can work on > 16 GBs of RAM with sub ms max latency :)

wmf• 17 years ago

Maybe if we're lucky Azul will go broke and Google will buy up the talent.

blasdel9• 17 years ago

It is intended to be absolutely safe: dereferencing an invalid pointer doesn't result in "undefined behavior" -- it results in your program halting every time. You can't accidentally succeed or allow for buffer overflows, which is all that's important (especially since Go doesn't have exceptions).

Hopefully they can rip off the design of GHC's amazing parallel garbage collector (it runs in its own pthread), and ditch the current mark-and-sweep.

telemachos2• 17 years ago

You have to love a project where the designers are sitting on irc fixing build bugs as they come in. (No, this isn't unique for opensource projects, but it's always great to see.)

Thanks for giving me something to do with all my free time for the next couple of weeks.

johnbender• 17 years ago

The main selling point is compile time.

Sweet...

euroclydon2• 17 years ago

"However, in Go, any type which provides the methods named in the interface may be treated as an implementation of the interface"

I've been wanting this in C# for a while now. Maybe one day...

AndrewDucker• 17 years ago

Yeah, I'd kill for Duck Typing in C#.

nailer19• 17 years ago

IronPython is duck typed and runs on the DLR. The same should be possible for a future C#...

euroclydon2• 17 years ago

I never knew what duck typing referred to. Thanks.

nikils2• 17 years ago

might be influenced from this http://www.vitanuova.com/inferno/papers/limbo.html

shabda3• 17 years ago

import "fmt"

what advantages this have over

import fmt

that a new language is choosing this sytax?

mncaudill• 17 years ago

It keeps the grammar simple, for one.

shabda3• 17 years ago

IMO, Thats no reason to make me type two extra chars each time I import.

demallien• 17 years ago

it allows module names with spaces/other special characters, for two.

akamaka2• 17 years ago

Could it have something to do with allowing unicode in the source?

praptak2• 17 years ago

It's a shorthand for:

import fmt "fmt"

Where the quoted fmt is an import path. Not making it a string literal would create a potential clash between characters used in filesystem paths and the Go parser.

swaroop• 17 years ago

I wonder if Go is what will "replace" Python internally at Google? ( w.r.t. http://news.ycombinator.com/item?id=933493 )

_cbsz2• 17 years ago

Apparently, Go style is to use tabs (see go-mode.el). This is not Pythonic.

jeremyw• 17 years ago

Yeah, this was most disappointing (I'm otherwise giddy). I hoped source tabs were dead and buried.

blasdel9• 17 years ago

There's also gofmt, which automatically applies the full house style using tabs.

If anyone can bring about a tab reniassance, it's bwk, ken, rob, rsc, and gri. These titans wrote your operating systems and your programming languages. They all wrote their own editors (not sure about gri) -- hell, the Go tests use ed!

Tabs have been lost in the wilderness for too long, it is about time they are reborn!

johnmw• 17 years ago

I am curious if someone could explain how a project in Go would compare to a project in Python or Ruby + a fancy array of custom C modules? This implies that it is possible to implement the nicer features of Go (for example concurrency support) in the modules supporting these languages. Or instead making kickarse fast compilers for higher level languages?

Is it too simplistic to say that at one end you have expressive languages designed for humans and at the other you have low level languages designed for speed?

edit: not meaning to insult humans who like fast languages.

makecheck• 17 years ago

It's a very promising language, and the tutorial's not bad.

The main quirk I see right now is that the syntax feels more like a shuffling of characters, rather than actually simplifying anything. For instance, they save parentheses in for loops (great idea!), yet suddenly I need a useless "func" before every function. It also seems like there needs to be a default type for arguments, either "int" or "string", for it to feel truly simplified.

A feature I would really, really like to see in a language is the notion of automatically printing lone strings, or automatically printf()-ing lone string format lists. It's a little irritating after all this time that one must "echo" or "print" or System.out.println() or fmf.Printf() everything (much less import standard libraries to do it), when it is such a common task.

dfarm2• 17 years ago

It does certainly smack of "Not-Invented-Here Syndrome," but compared to the Dilbert like project management culture of my job I rather admire it.

I've just been getting in to generic programming with C++ (Stepanov's new book, awesome). So the fact that they don't have templates or operator overloading (the two critical pieces to STL style) was a bit of a downer. That being said the interfaces concept reminded me of generic functions (in CLOS). Sounds pathetic but I probably will give it a try just because it's from Google -- but I'm quite skeptical. With the really interesting work going on with Haskell, Clojure, Fortress, etc. it seems a shame to stay so close to C.

jacquesm10• 17 years ago

Have a look at some of the names behind it and see if you stick to your 'Not Invented Here'.

ionfish• 17 years ago

"Reinventing the wheel we invented earlier"?

(That's not a criticism—language design shouldn't always be in the revolutionary mode.)

dfarm2• 17 years ago

I don't think you understand what "Not invented here syndrome" means. See: http://en.wikipedia.org/wiki/Not_Invented_Here

What I mean is that Google always has to do things the Google way whether that's an improvement or not. Whether this turns out to be a valuable contribution to the field will play out over the next few years. We'll see. But a company that makes their own DB, their own phone, their own programming languages, etc. is pretty much a prototypical "not invented here" company.

Slashed2• 17 years ago

This new programming language from Google is quite a buzz on the web now. To be honest, I always got excited about tools that promised to speed-up and simplify development. But most of the time, those trendy tools are just even more time wasters. One language I find to be [almost] ideal is Scala. I believe that Scala will replace Java. Also, writing code for JVM gives you a huge advantage - you can create OSGi bundles. IMO, OSGi framework is the most mature module system. So I think that there's nothing good about this language except that it's created by Google.

gchpaco2• 17 years ago

Some of the style reminds me of Alef from Plan 9; in particular the communication channels, but also a fair chunk of the syntax.

wendroid2• 17 years ago

That will be because that's what it's based on.

FreeRadical• 17 years ago

Can someone give me some examples of where this might be used (and to do what)?

blasdel9• 17 years ago

  hg log --rev 0:4 --template '{date|isodate}\t{author}\n\t\t\t\t{desc|firstline}\n\n'
  1972-07-18 19:05 -0500	Brian Kernighan <bwk>
  				hello, world
  
  1974-01-20 01:02 -0400	Brian Kernighan <bwk>
  				convert to C
  
  1988-04-01 02:02 -0500	Brian Kernighan <research!bwk>
  				convert to Draft-Proposed ANSI C
  
  1988-04-01 02:03 -0500	Brian Kernighan <bwk@research.att.com>
  				last-minute fix: convert to ANSI C
  
  2008-03-02 20:47 -0800	Robert Griesemer <gri@golang.org>
  				Go spec starting point.

antirez2• 17 years ago

Ok, a language for engineers not for people that want to be truly expressive with the code they write. This is my impression when looking at the examples, they have something odd in the way the code looks like.

Now from the technical point of view, there are a few interesting points like the threading approach, but in general the language does not appear to provide great new ideas, nor to be a language where consolidated ideas are put together very elegantly.

It seems like that for system programming the "better C" that one could need is different than this, while for general programming what is really missing is something like Ruby and Python but made as fast as C and with low memory usage by mean of doing the right sacrifices to the language, but still retain most of the abstraction level.

Go is not as powerful as C++, not as high level as Java, not as raw as C for low level stuff. I don't like it.

blasdel9• 17 years ago

Neither C++ nor Java implement Hoare's CSP, or have any primitives fundamental enough to implement them in terms of.

Go is exactly what you're asking for in your third paragraph -- it seems you were blinded by the curly braces...

antirez2• 17 years ago

Yes CSP is a good abstraction that Go implements, but programming languages are not just a matter of features like call-cc, tail-calls, closures and so on. This are very important things, but it's not enough.

fauigerzigerk• 17 years ago

The language breaks with a tradition established in the past 15 years, which says that every new language must be at least a few hundered times slower and less memory efficient than everything that came before it (I know there are exceptions)

Go is simple and fast. It has garbage collection, it facilitates parallel programming and it does away with lots of useless OO cruft. At first sight I like it.

jgrahamc• 17 years ago

For my doctorate I did everything in CSP and Occam, so it's wonderful to see the channel communication methods defined for this language.

http://golang.org/doc/go_mem.html

If you do things via synchronized channel communication then it eliminates (mostly) the need to spend all your time with mutexes and the like because the channel sync does that work for you. At the beginning it'll probably take people time to get there heads around it, but my recollection was that it was a game changer once you understood how the synchronization worked.

olaf3• 17 years ago

I would like to see Google "eat their own dogfood" before I invest in Go. Or do they already use Go in production?

uriel2• 17 years ago

goglang.org runs Go.

olaf3• 17 years ago

I meant real business like GMail or similar.

Semiapies2• 17 years ago

I agree as to "investing", but as to experimenting with and trying out, I don't see why not. And the site running on Go seems to be puttering along nicely...

c00p3r2• 17 years ago

And finally C 2.0 sponsored by Google. Good move against Apple.

It (go + native client) might become a next major platform. Just think about low-level (and easier to code than C[++]) Javascript alternative.

rsaarelm• 17 years ago

A lot of this looks very nice. One very simple thing I noticed was missing was a binary integer literal syntax. With a systems programming language, it would be nice to be able to write bit sets without shifts sometimes. Just use the obvious '0b10100101' syntax.

Also, I think the zero-prefix convention for octal literals is very confusing and possibly error-prone, but it seems to be used even in modern programming languages. Is there a good reason to keep using '0765' instead of something '0o765'?

The Go Programming Language

Related Stories