Go is Like Better C, Mostly

The Go programming language is was on my (long) list of things to look into for quite some time now. Recently, at last, I had the opportunity to go through the most part of a comprehensive tour of Go from the official website, as well as write few bits of some Go code by myself.

Today I’d like to recap on some of my impressions. You can treat it as “unboxing” of the Go language, much like when people post movies of their first hands-on experiences with new devices. Except, it will be just text – I’m not cool enough to do videos yet ;)

Some trivia

We all like to put stuff into our various mental buckets, so let’s do that with Go too.

Go is a compiled, statically typed programming language that runs directly on the hardware, without any underlying virtual machine or other bytecode-based runtime. That sounds good from the speed viewpoint and indeed, Go comes close to C in raw performance of equivalent programs.

Syntax of Go is C-like, at least in the fact that it’s using curly braces to delimit blocks of code. Some visual clutter is intentionally omitted, though. Semicolons are optional, for example, and idiomatic Go code omits them at all times.
But more surprisingly, parentheses around if and for conditions are straight out forbidden. As a result, it’s mandatory to use curly braces even for blocks that span just one line:

If you’re familiar with reasoning that suggests doing that in other C-like languages, you shouldn’t have much problems adapting to this requirement.

No-fuss static typing

Go is type-safe and requires all variables to be declared prior to use. For that it provides very nice sugar in the form of := operator, coupled with automatic type inference:

But of course, function arguments and return values have to be explicitly typed. Coming from C/C++/Java/etc. background, those type declarations might look weird at first, for they place the type after the name:

As you can see, this also results in putting return type at the end of function declarations – something that e.g. C++ also started to permit.

But shorthand variable declarations are not the only way Go improves upon traditional idioms of static typing. Its interfaces are one of the better known features here. They essentially offer the support for duck typing (known from Python, among others) in a compiled language.
The trick is that objects do not specify which interfaces they implement: it’s just apparent by their methods. We can, however, state what interfaces we require for our parameters and variables, and those constraints will be enforced by the compiler. Essentially, this allows for accepting arbitrary values, as long as they “quack like a duck”, while retaining the overall type safety.

As an example, we can have a function that accepts a very general io.Writer:

and use it with anything that looks like something you could write into: file objects, networked streams, gzipped HTTP responses, and so on. Those objects won’t have to declare or even know about io.Writer; it’s sufficient that they implement a proper Write method.

Pointers on steroids

Talking about objects and interfaces sounds a bit abstract, but we shall not forget that Go is not a very high level language. You still have pointers here like in C, with the distinction between passing an object by address and copying it by value like in C++. Those two things are greatly simplified and made less error prone, however.

First, you don’t need to remember all the time whether you interact with object directly or through a pointer. There’s no -> (“arrow”) operator in Go, so you just use dot (.) for either. This makes it much easier to change the type of variable (add or remove *) if there’s need.

Second, most common uses for pointers from C (especially pointer arithmetic) are handled by dedicated language mechanism. Strings, for example, are distinct type with syntactic support and not just arrays of chars, neither a standard library class like in C++. Arrays (called slices) are also well supported, including automatic reallocation based on capacity, with the option of reserving the exact amount of memory beforehand.

Finally, the common problems with pointer aliasing don’t really exist in Go. Constraints on pointer arithmetic (i.e. prohibiting it outright) mean that compiler is able to track how each and every object may be used throughout the program. As a side effect, it can also prevent some segmentation faults, caused by things like local pointers going out of scope:

The i variable here (or more likely: the whole stack frame) will have been preserved on heap when function ends, so the pointer does not become immediately invalid.

Packages!

If you ever coded a bit in some of the newer languages, then coming to C or C++ you will definitely notice (and complain about) one thing: lack of proper package management. This is an indirect result of the header/implementation division and the reliance on #include‘ing header files as means of specifying dependencies. Actually, #includes are not even that: they work only for compiler and not linker, and are in some sense abused when working with precompiled headers.

What about Go?… Turns out it does the right thing. There are no separate header and implementation units, only modules (.go files). Unless you are using GCC frontend or interfacing with C code, the compiler itself is also unified.

But most importantly, there are packages and normal import statements. You can have qualified and unqualified imports, and you can alias things you’re importing into different names. Packages themselves are based on directory structure rooted in $GOROOT, much like e.g. Python ones are stored under $PYTHONPATH.

The only thing you can want at this point is the equivalent of virtualenv. Note that it’s not as critical as in interpreted languages: standalone compiled binaries do not have dependency problems, after all. But it’s still a nice thing to have for development. So far, people seem to be using their own solutions here.

Of Languages and Keywords

I recently had a discussion with a co-worker about feasibility of using anonymous functions in Python. We happen to overuse them quite a bit and this is not something I’m particularly fond of. For me lambdas in Python are looking pretty weird and thus I prefer to use them sparingly. I wasn’t entirely sure why is it so – given that I’m quite a fan of functional programming paradigm – until I noticed a seemingly insignificant fact.

Namely: the lambda keyword is long. With six letters, it is among the longer keywords in Python 2.x, tied with return, import and global, and beaten only by continue and finally. Quite likely, this is what causes lambdas in Python to stand out and require additional mental resources to process (assuming we’re comfortable enough with the very idea of anonymous functions). The long lambda keyword seems slightly out of place because, in general, Python keywords are short.

Or are they?… Thinking about this, I’ve got an idea of comparing the average length of keywords from different programming languages. I didn’t really anticipate what kind of information would be exposed by applying such a metric but it seemed like a fun exercise. And it surely was; also, the results might provoke a thought or two.

Here they are:

The newest incarnation of C++ seems to be losing badly in this competition, followed by C#. On the other side of the spectrum, Go and Python seem to be deliberately designed to avoid keyword bloat as much as possible. Java is somewhere in between when it comes to sheer numbers of keywords but their average length is definitely on the long side. This could very well be one of the reasons for the perceived verbosity of the language.

For those interested, the exact data and code I used to obtain these statistics are in this gist.

pyduck – biblioteka do interfejsów w Pythonie

Czas pochwalić się swoim nowym dziełem. Nie jest ono bardzo imponujące ani specjalnie duże. Mam jednak nadzieję, że będzie ono przydatne dla tego (wąskiego) grona odbiorców, do którego jest skierowane.

Mam tu na myśli niewielką biblioteką do Pythona, która ma na celu poprawienie użyteczności jednej z głównych, ideowych cech języka – tak zwanego typowania kaczkowego (duck typing). Geneza tego terminu jest oczywiście wielce intrygująca, ale założenie jest proste. Zamiast czynić obietnice i jawnie deklarować implementowane interfejsy, obiekty w Pythonie “po prostu są” i zwykle próbują być od razu używane do założonych celów. Jeśli okażą się niekompatybilne (np. nie posiadają żądanej metody), wtedy oczywiście rzucany jest wyjątek. Pythonowska praktyka polega więc na przechodzeniu do rzeczy bez zbędnych ceregieli i obsłudze ewentualnych błędów.

Ma to rzecz jasna swoje zalety, ma też wady, a czasami może również rodzić problemy, jeśli błąd spowodowany niekompatybilnością obiektu ujawni się za późno. Z drugiej strony brak konieczności jawnego specyfikowania implementowanych interfejsów to spora zaleta. Najlepiej więc byłoby jakoś połączyć te dwa światy i umożliwić wcześniejsze sprawdzenie możliwości obiektu…

Jak można się pewnie domyślić, to właśnie próbuje umożliwić mój framework, noszący wdzięczną nazwę pyduck. Dodaje on do Pythona mechanizm interfejsów bardzo podobny do tego, który obecny jest w języku Go. Najważniejszą jego cechą jest właśnie fakt, że w konkretnych typach interfejsy są implementowane niejako automatycznie – wystarczy, że mają one odpowiednie metody. Samo sprawdzenie, czy obiekt implementuje dany interfejs polega zaś na faktycznym zaglądnięciu w listę jego metod, a nie weryfikacji jakichś jawnych deklaracji.

Inaczej mówiąc, nie czynimy tutaj żadnych obietnic odnośnie obiektu, ale wciąż mamy możliwość kontroli, czy nasze wymagania są spełnione. Najlepiej ilustruje to oczywiście konkretny przykład:

Zaznaczamy w nim, że metoda Canvas.draw_object spodziewa się obiektu zgodnego z interfejsem Drawable. Jest on zdefiniowany wyżej jako posiadający metody get_bounds i draw. Sprawdzenie, czy rzeczywisty argument funkcji spełnia nasze wymogi, zostanie wykonane przez dekorator @expects. Zweryfikuje on obecność i sygnatury metod wspomnianych metod.
Dzięki temu będziemy mogli być pewni, że mamy do czynienia z obiektem, który rzeczywiście potrafi się narysować. Jego konkretna klasa nie będzie musiała natomiast nic wiedzieć na temat interfejsu Drawable ani jawnie deklarować jego wykorzystania.

Po więcej informacji zapraszam oczywiście na stronę projektu na GitHubie. Ewentualnie można też od razu zainstalować paczkę, np. poprzez easy_install:

A ponieważ wszystko open source jest zawsze wersją rozwojową, nie muszę chyba wspominać, że z chęcią witam pull requesty z usprawnieniami i poprawkami :>