Suites, scopes, and lifetimes

To my post Why Python doesn't need blocks, julien tayon replied with a comment about a completely different meaning of blocks, which I think leads to some interesting points, even if it's irrelevant to that post.

Terminology

Like many words in computing, "block" is overloaded with multiple, contradictory meanings.

Ruby has a syntactic feature that it calls "blocks," a way to define special-purpose anonymous callables, which it calls "procs". Objective-C (and Apple-extended C) borrowed that feature, and used the word "block" to refer both to the syntax for defining these things, and for the things so defined.

But meanwhile, there's a completely separate concept in most imperative languages: a sequence of statements set off by braces, indentation, begin/end, etc., that can be used syntactically like a single statement. And "block" is a pretty common word to refer to this idea, but I'll use the term "suite", as used by Python's reference documentation, to avoid confusion.

Julien's reply

Scope, what is it good for?

well scopes (not blocks) are for limiting the existence of a variable so that it does not clubber the namespace later and make side effects.

This is only half of what scopes do, and not the most interesting half.

Scopes are closures—the variables defined in a scope are only available inside that scope. But in a language with lexical scoping and first-class functions, this means you can capture those variables in a lexical closure and pass it around. That isn't possible in C, but in Python (and many other languages), you can write this:

def make_adder(addend):
def adder(augend):
return augend+addend
return adder

adder_of_5 = make_adder(5)
adder_of_10 = make_adder(10)

(And, at the risk of increasing the confusion again, it's this use of scopes that Ruby's blocks were invented for—unlike Python, Ruby functions are not lexically scoped, so they needed something else that was.)

I don't want to give a full tutorial on closures, how cool they are, and the effects of alternative scoping rules, etc.; if you don't already know all that, you'll find better versions on Google (or probably even Wikipedia).

It is also useful for enforcing the locality of variable (which is very usefull for triggering CPU cache optmization but since python variables are boxed it is not very useful in python).

I'll ignore the micro-optimization stuff here, because it really isn't relevant.

Suites and scopes

Python is lacking of blocks, I miss strictures from Perl.

Python is not lacking either suites (blocks) or scopes.

The difference between Python and the C family of languages is that not every suite defines a scope; basically, it's just function and class definitions.

So, why not have every suite define a scope?

First, there's not much benefit (except in the case of a few unusual languages like C++, which I'll get to in the next section). Functions can bind scopes, but for statements can't. So, all you get out of a scoped for statement is the ability to (usually accidentally) shadow outer variables. If you look inside the code generated by a C compiler, a function's stack frame has room for all of the variables defined in the deepest scope in the function, just like in Python.

And there is a cost. Besides making the language harder to understand, compile, and interpret, it makes it harder to debug your code. For example, in this function:

int foo() {
int bar = 1;
// ...
if (1) {
int bar = 2;
// ...
}
// ...
return bar;
}

… there are two different variables, both named bar, where the inner one shadows the outer. Of course in real life, this would probably be a 120-line function, where the two definitions were 58 lines apart and couldn't even fit on the screen together. This makes it harder for a reader to understand the code. And it means a debugger needs some way to distinguish "not that bar, the other bar in the same function". And it means the debugger's interface needs some way to let you distinguish them in your commands.

In C, there are (at least historical) reasons why long, complex functions are worth having, and occasionally maybe it's even worth reusing the same variable name 58 lines later because it's the locally most-readable name. (That being said, in some cases, compilers will warn you about doing so…)

But in most newer languages—including Python—there's no reason not to break your code up into smaller pieces. If your inner code needs to use the name bar for readability, factor it out into a separate function. You can even keep that function inline if you want (gaining the added advantage over C that you can explicitly control which variables you do and don't want to share with the parent scope):

def foo():
bar = 1
# ...
def inner():
bar = 2
# ...
inner()
# ...
return bar

So, you can have a new scope whenever you want it, but you don't get one when you don't need it. You don't accidentally shadow variables, but you can do so when you need to. And so on.

RAII

C++, like Python, is designed to encourage small functions, but it still makes extensive use of suite scopes. In fact, you'll even often see "bare suites" created just for scoping purposes:

void foo(shared_ptr<thing_t> thing) {
immutable_stuff(thing);
{
lock_t lock(thing->mutex);
mutating_stuff(thing);
}
more_immutable_stuff(thing);
}

This is a feature that C++ calls Resource Acquisition Is Initialization.

In C++, everything, even class types, gets allocated on the stack, and destroyed when it goes out of scope. If you want something to live longer, you have to use the new operator to allocate it on the heap, use a pointer to it, and delete it when you're done. If you need to allocate it on the heap, but still want it to go away at the end of the scope, you can allocate a smart pointer (the stdlib comes with various different variations) that itself lives on the stack, goes away at the end of the scope, and deletes the object.

Building on that, each class can have a destructor that gets called right before the object gets destroyed. So, if you acquire a resource in your constructor, and release it in your destructor, you can tie the resource's lifetime to your object's lifetime, which you can tie to the scope.

The problem with this feature is that it relies on manual memory management. In most other current languages, objects are allocated on the heap, and automatically garbage collected after you no longer refer to them. Since the garbage collector isn't tied to any scope, you can't rely on destructors to tie resource lifetimes to scopes. Instead, you have to do it using something like Java's try/finally statement:

void foo(Thing thing) {
immutable_stuff(thing);
try {
lock(thing->mutex);
mutating_stuff(thing);
} finally {
unlock(thing->mutex);
}
more_immutable_stuff(thing);
}

In that code, the extra scope created by the try suite is obviously not helping anything. Without RAII, suite scopes aren't useful.

Special-purpose suites

Many C++-family languages have a special kind of statement that wraps this up lock-a-mutex-for-the-duration-of-a-suite use case:

void foo(Thing thing) {
immutable_stuff(thing);
synchronized(thing->mutex) {
mutating_stuff(thing);
}
more_immutable_stuff(thing);
}

But that only works for locks; it doesn't work for files or large chunks of memory or anything else, unless the language has a special statement for each kind of resource. RAII effectively exposes a try/finally contexts in a way that's accessible at the language level, so you (or the stdlib, or a third-party library) can write any number of "synchronized"-like wrappers for every kind of object you want.

Python has a very similar, but different, solution to the same problem, as Julien acknowledges:

Context managers

And python with context manager has finally began to understand the wisdom in scopes:

with open("/tmp/whatever") as f: do some_thing_of(f)

# f does not exists here, and file is closed \o/

I'm not sure what's meant by "finally" here, since context managers have been part of Python since 2.5 (see PEP 343), and were under discussion for 3 years before being implemented.

Also, it's not true that f does not exist after the with statement. It still exists; you can print it out and see that it's a perfectly valid TextIOWrapper around a closed file.

And this is the key. With statements are not about scopes; they're a way of managing resource lifetimes independently of the scopes of any variables that may reference them. You get all of the advantages of RAII without needing the otherwise-useless scoped suites.

(The fact that a with statement has a suite, and so does a function definition or other scoped statement, is really only marginally relevant here; all compound statements have suites, most of them don't have scopes, and most of them have nothing to do with resource lifetime.)

On the other hand, there is an advantage of RAII over context managers, and most other languages' solutions. RAII effectively lets you turn any scope into a context manager. If you already have a suite whose scope is "close enough" to the desired lifetime of the resource, you can just drop an RAII object into that suite, without needing to create an enclosed suite. Whether that benefit is worth the cost of explicit memory management is a question for debate.

It's been more than a decade since Typical Programmer Greg Jorgensen taught the word about Abject-Oriented Programming.

Much of what he said still applies, but other things have changed. Languages in the Abject-Oriented space have been borrowing ideas from another paradigm entirely—and then everyone realized that languages like Python, Ruby, and JavaScript had been doing it for years and just hadn't noticed (because these languages do not require you to declare what you're doing, or even to know what you're doing). Meanwhile, new hybrid languages borrow freely from both paradigms.

This other paradigm—which is actually older, but was largely constrained to university basements until recent years—is called Functional Addiction.

A Functional Addict is someone who regularly gets higher-order—sometimes they may even exhibit dependent types—but still manages to retain a job.

Retaining a job is of course the goal of all programming. This is why some of these new hybrid languages, like Rust, check all borrowing, from both paradigms, so extensively that you can make regular progress for months without ever successfully compiling your code, and your managers will appreciate that progress. After all, once it does compile, it will definitely work.

Closures

It's long been known that Closures are dual to Encapsulation.

As Abject-Oriented Programming explained, Encapsulation involves making all of your variables public, and ideally global, to let the rest of the code decide what should and shouldn't be private.

Closures, by contrast, are a way of referring to variables from outer scopes. And there is no scope more outer than global.

Immutability

One of the reasons Functional Addiction has become popular in recent years is that to truly take advantage of multi-core systems, you need immutable data, sometimes also called persistent data.

Instead of mutating a function to fix a bug, you should always make a new copy of that function. For example:

function getCustName(custID)
{
    custRec = readFromDB("customer", custID);
    fullname = custRec[1] + ' ' + custRec[2];
    return fullname;
}

When you discover that you actually wanted fields 2 and 3 rather than 1 and 2, it might be tempting to mutate the state of this function. But doing so is dangerous. The right answer is to make a copy, and then try to remember to use the copy instead of the original:

function getCustName(custID)
{
    custRec = readFromDB("customer", custID);
    fullname = custRec[1] + ' ' + custRec[2];
    return fullname;
}

function getCustName2(custID)
{
    custRec = readFromDB("customer", custID);
    fullname = custRec[2] + ' ' + custRec[3];
    return fullname;
}

This means anyone still using the original function can continue to reference the old code, but as soon as it's no longer needed, it will be automatically garbage collected. (Automatic garbage collection isn't free, but it can be outsourced cheaply.)

Higher-Order Functions

In traditional Abject-Oriented Programming, you are required to give each function a name. But over time, the name of the function may drift away from what it actually does, making it as misleading as comments. Experience has shown that people will only keep once copy of their information up to date, and the CHANGES.TXT file is the right place for that.

Higher-Order Functions can solve this problem:

function []Functions = [
    lambda(custID) {
        custRec = readFromDB("customer", custID);
        fullname = custRec[1] + ' ' + custRec[2];
        return fullname;
    },
    lambda(custID) {
        custRec = readFromDB("customer", custID);
        fullname = custRec[2] + ' ' + custRec[3];
        return fullname;
    },
]

Now you can refer to this functions by order, so there's no need for names.

Parametric Polymorphism

Traditional languages offer Abject-Oriented Polymorphism and Ad-Hoc Polymorphism (also known as Overloading), but better languages also offer Parametric Polymorphism.

The key to Parametric Polymorphism is that the type of the output can be determined from the type of the inputs via Algebra. For example:

function getCustData(custId, x)
{
    if (x == int(x)) {
        custRec = readFromDB("customer", custId);
        fullname = custRec[1] + ' ' + custRec[2];
        return int(fullname);
    } else if (x.real == 0) {
        custRec = readFromDB("customer", custId);
        fullname = custRec[1] + ' ' + custRec[2];
        return double(fullname);
    } else {
        custRec = readFromDB("customer", custId);
        fullname = custRec[1] + ' ' + custRec[2];
        return complex(fullname);
    }
}

Notice that we've called the variable x. This is how you know you're using Algebraic Data Types. The names y, z, and sometimes w are also Algebraic.

Type Inference

Languages that enable Functional Addiction often feature Type Inference. This means that the compiler can infer your typing without you having to be explicit:

function getCustName(custID)
{
    // WARNING: Make sure the DB is locked here or
    custRec = readFromDB("customer", custID);
    fullname = custRec[1] + ' ' + custRec[2];
    return fullname;
}

We didn't specify what will happen if the DB is not locked. And that's fine, because the compiler will figure it out and insert code that corrupts the data, without us needing to tell it to!

By contrast, most Abject-Oriented languages are either nominally typed—meaning that you give names to all of your types instead of meanings—or dynamically typed—meaning that your variables are all unique individuals that can accomplish anything if they try.

Memoization

Memoization means caching the results of a function call:

function getCustName(custID)
{
    if (custID == 3) { return "John Smith"; }
    custRec = readFromDB("customer", custID);
    fullname = custRec[1] + ' ' + custRec[2];
    return fullname;
}

Non-Strictness

Non-Strictness is often confused with Laziness, but in fact Laziness is just one kind of Non-Strictness. Here's an example that compares two different forms of Non-Strictness:

/****************************************
*
* TO DO:
*
* get tax rate for the customer state
* eventually from some table
*
****************************************/
// function lazyTaxRate(custId) {}

function callByNameTextRate(custId)
{
    /****************************************
    *
    * TO DO:
    *
    * get tax rate for the customer state
    * eventually from some table
    *
    ****************************************/
}

Both are Non-Strict, but the second one forces the compiler to actually compile the function just so we can Call it By Name. This causes code bloat. The Lazy version will be smaller and faster. Plus, Lazy programming allows us to create infinite recursion without making the program hang:

/****************************************
*
* TO DO:
*
* get tax rate for the customer state
* eventually from some table
*
****************************************/
// function lazyTaxRateRecursive(custId) { lazyTaxRateRecursive(custId); }

Laziness is often combined with Memoization:

function getCustName(custID)
{
    // if (custID == 3) { return "John Smith"; }
    custRec = readFromDB("customer", custID);
    fullname = custRec[1] + ' ' + custRec[2];
    return fullname;
}

Outside the world of Functional Addicts, this same technique is often called Test-Driven Development. If enough tests can be embedded in the code to achieve 100% coverage, or at least a decent amount, your code is guaranteed to be safe. But because the tests are not compiled and executed in the normal run, or indeed ever, they don't affect performance or correctness.

Conclusion

Many people claim that the days of Abject-Oriented Programming are over. But this is pure hype. Functional Addiction and Abject Orientation are not actually at odds with each other, but instead complement each other.