Date: Tue Dec 29 2015 16:00:00 GMT-0800 (PST)

Updated: Wed Apr 27 2016 17:00:00 GMT-0700 (PDT)

This is What Programming Language You Should Learn

Aspiring coders tend to ask, “What programming language should I start with?” It’s not a bad question, but it’s not really the right one to ask. Why?

When newbs ask for language recommendations, they typically get one of two things. The first is something to the effect of “It doesn’t matter. Choose one and stick with it.” It’s not a bad answer—in fact it’s correct—but it’s not really the right one to offer beginners1.

The second is someone’s personal favourite language, which is often the only one they’ve ever used. This is a bad answer and if you do this, you should feel bad.

This has got to be immensely frustrating for newbs who just want to get the ball rolling. Though language choice is of little consequence, people keep asking about it. I attempt to answer it, to the extent it can be answered, not with a glib response like “JavaScript”2, but rather with the questions you should be asking and a thorough overview of the landscape of coding today.

It’s more important for a budding coder to gain momentum and have fun doing it than it is to start in the “right” place. However, as time goes on and one’s skills improve, one will find that certain patterns of knowledge complement each other readily, enabling a higher overall problem-solving potential. I’ll try to describe and delineate those patterns—to the extent that I (a generalist, game developer, and open-source fan) understand them.

This article is geared towards people who have never (or rarely) programmed before but who are considering a career or otherwise substantial investment of time into learning to code. I hope it may be useful to those who want to pick up coding for shiggles or perhaps to automate a part of their job, but it’s not my primary focus.

Without further ado, here’s a bigass list of…

Crap Every Programmer Must Learn

Briefly, and in no particular order:

  1. Languages
  2. Concepts
  3. Libraries
    • Modules
    • Frameworks
    • Software Development Kits
  4. Tools
  5. Soft Skills

You will need to learn each major category in parallel (there’s simply no way not to), but within each category I find it best—if possible—to focus on one skill at a time, from a new subcategory each time. We’ll start with the reason you clicked here in the first place.

Languages

We cannot compare different languages without talking about the idea of high-level and low-level programming. Programming is built around the idea of abstractions. For example, instead of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, I can just say “the numbers 1 through 10”, and the machine—as well as any first-grader—will understand what I mean. We can make abstractions upon abstractions, and in so doing, we venture into higher levels of programming. Briefly, like a house is built first with a foundation, then a wooden frame (or perhaps brickwork), then walls and finally paint, shingles, and doodads, so too we can build programming languages on top of one another, allowing us to do bigger and better things with less and less code. 4 High-level languages like JavaScript allow us to push the boundaries of computing, but they cannot exist without low-level ones like C.

Kind of like the sciences

Programming languages come in four basic flavors: system, scripting, managed, and data. Oh, there are most certainly others, but these are the ones you’ll find yourself dealing with from day to day, for reasons that are beyond the scope of this article.

Systems Programming

The lowest-level type of programming as we typically think of it. Fortran and C are the major examples here, although you won’t find Fortran in modern usage, except in highly scientific domains. C++ is a bit of a beast in a category all its own, but may be considered a systems language.

Learning a systems language will empower you to write software that runs not only in browsers or PCs, but in toasters, cars, wearables, and devices that have yet to be invented.

Programming in a systems language is not easy. However, learning to program in one is. The relative simplicity of the lower levels means there are fewer nasty surprises outside of your control, but you will wind up writing your own nasty surprises as your projects grow in complexity.

There’s been an uptick in high-level native languages like Haskell, Rust or Go, which compile down to machine code like C or C++, but operate at a level of abstraction more typical of scripting languages. Along with next-generation compiler tech like LLVM, such languages offer an enticing blend of productivity and performance that could give managed languages a run for their money in the coming years.

Managed Languages

Managed languages are those that are designed to run on a virtual machine, rather than directly on your CPU. They are so named because they manage some of the more mundane details of execution (most notably memory allocation), freeing you, the programmer, from fretting about it. There are two such machines in common use today: the Java Virtual Machine which of course powers Java and its family, and the Common Language Infrastructure, which powers C# and its ilk.

Managed languages are just about as general-purpose as they come. Business and backend software overwhelmingly favor managed languages as they strike a nice balance between performance and coder productivity—that is, between speed of execution and speed of delivery, or if you like, between CPU time and developer time. The tradeoff between machine and human time is an important consideration in modern software development that’s rarely taught in the classroom (and beyond our scope here). However, the popularity of managed languages in industry means that a formal education in computing will almost certainly introduce you to them.

Scripting Languages

Scripting languages are generally the highest level you’ll encounter as a typical programmer, and thus do the most work with the smallest amount of code. Python, Perl, Ruby, and Lua are some popular examples. Typically, scripting languages are interpreted in real-time, meaning that there’s no need to compile them before running—just change your code and go. They’re particularly well-suited to automating human work, such as file renames, sending email, or web crawling, as well as to so-called “glue code” that ties different systems together.

JavaScript is a special case of scripting language, in that it was specifically designed to run in the browser, in the context of a web page. It eventually came to reign victorious as the de facto lingua franca of the web. With the advent of Node, it extended into the realm of “traditional” desktop scripting languages, but its origins in the web give it historical baggage that confounds developers to this day.

Although flexible and powerful, scripting languages aren’t very performant and may take orders of magnitude more CPU time to do the same thing as the same program written in C. Now, often, that doesn’t matter. But sometimes it does.

Data Languages

Pedants will point out that data languages aren’t, technically, languages—but whatever they are, they’re hella useful and you should get familiar with several and be an expert at your favourites.

XML, JSON, and Yaml are examples of data interchange formats. They model “stuff” that can be loaded into your program in many different ways. Each has its strengths and weaknesses, but all are easy to use from any programming language and will save you plenty of headaches over rolling your own format like my dumb ass once did.

HTML, Markdown, LaTeX are markup languages, a special form of data language geared towards writing documents. HTML, of course, is the de facto language of the web and it’s hard to be a modern techie without getting at least a little familiar with it. LaTeX on the other hand is geared towards printable documents and is commonly used to produce PDFs. Markdown is a more minimalist format modeled after email conventions but it’s become popular with communities like GitHub, Reddit and StackOverflow adopting it for its ease of use. This article was written in Markdown.

By the way, you can combine markup languages with programming languages if you want to describe a bit of code in detail. That’s called literate programming and I think it’s mad cool. Here’s a small literate program I wrote..

Concepts

This is the stuff they teach you in college. The reason it’s taught in college is because, if it weren’t, it would take many years to emerge “naturally” as abstractions upon concrete lessons learned. The entire space of CS concepts is rather vague, with many grey areas. In my eye, concepts come in two main flavors:

A programming paradigm is a set of both control and data constructs that work reasonably well together and are frequently found snuggling together in the form of a particular language. Notably, the functional, procedural, and declarative paradigms are highly influential in modern programming. Object Oriented Programming gets all the attention but it’s really just a particular formulation of those three. Mastering them first will help you understand why OOP is as effective as it is. For example: the very definition of a class is data plus controlling code.

Coding concepts are inextricably tied to programming languages that embody them, and so it’s not feasible to study them “purely”, as divorced from languages. Still, you should pay special attention to the concepts exposed in your language of choice, and challenge yourself on occasion to bend a language into using a concept it wasn’t designed for.

Control Concepts

Every (electronic) computer is ultimately a circuit with electrons shooting through it, really really fast. By carefully designing this circuit, we can control exactly how those electrons behave—choreographing millions of electric pulses into an intricate dance we call software.

Let’s take a modest processor operating at 1 GHz. This means it goes through one billion clock cycles per second. Each clock cycle (i.e. every nanosecond) this processor executes a single instruction and does the thing that the instruction says to do. Instructions might be be any of several dozen to several hundred basic operations, depending on the processor’s instruction set architecture. A single instruction might say “multiply these two bytes together and put the result in that byte”, or “set this byte to zero”.5.

Usually, instructions are processed one-by-one in the order they appear, kind of like following a cookie recipe. The real magic comes into play with special instructions that tell the process to branch based on certain conditions, which turns our simple recipe into a nonlinear choose-your-own-adventure novel where something different might happen every time.

In most programming languages, branching takes the form of the if statement—and occasionally, the notorious goto. Abstracting upon the humble if gives us more powerful ideas like iteration (doing the same thing over and over), recursion (code that calls itself), and others.

GOTO considered fatal

Data Concepts

Like code, digital data is nothing but heaps of ones and zeroes. What makes it interesting is how it is organized and interpreted.

The pointer is among the most fundamental data concepts but is notoriously difficult to wrap your head around. Pointers are sometimes called references5. Pointers are, simply, data whose purpose is to point to something else. This is called indirection and is a simple form of abstraction.

castle

Hyperlinks are a form of pointer you’re probably familiar with: http://www.google.com is useless data, but what is useful is the search engine served up by pointing a browser at it (pun intended). Similarly, you may have some bit of data 0xd34dbe3f that doesn’t do much until you dereference it, fetching the memory stored at that location, which might be a picture of a cat or something.

Composition (not the musical kind) is the idea of bundling related data together. Composition implies a “has-a” relationship; that is, that one thing has a different thing, in the sense that your wallet has a debit card, a credit card, and an ID.

struct Wallet {
  Card credit_card;
  Card debit_card;
  Card driver_licence;
  int dollars;
}

Composition lets us write programs that model real stuff with grace. Without it, we’d be limited to just numbers and letters, perpetually reliving the first grade. Yech.

Containers are an evolved form of composition where a data structure has an arbitrary amount of the same type of stuff. The order of that stuff may matter (list, array, tree), or it may not (set, map). When the term “data structures” is used in an academic context, it usually means containers.

Libraries

Libraries come in many shapes and sizes. At one end of the spectrum we have the module, which is just a small collection of related “stuff” which serves some need. Then we have the “framework”, which solves a very particular problem with a large amount of functionality. Frameworks typically make it easy to see fast results at the expense of learning curve, rigidity, and occasionally difficult debugging.

An SDK is a set of software including libraries but might contain compilers, drivers, tools, documentation and/or examples, typically for unusual hardware or other “platforms”.

Mastering libraries is, relatively speaking, a poor investment in the long term because they rise and fall as technology progresses—especially the React-Angular-Phaser-Kitchen-Sink Du Jour. That said, you’ll want to practice getting familiar with new libraries quickly and efficiently, since doing so will allow you to benefit from prior efforts rather than rolling your own every single time. Sometimes rolling your own is the right thing to do, either because you need to do something specific and special, because you know you can do it better than the existing solutions, or just to learn. Sometimes it’s not.

Learning to judge which route to go is a crucial skill that comes from lots of trial, error, and failure. I still suck at it.

A few of my favorite libraries, in no particular order, include SDL, momentjs, Joda-Time and Cordova. These offerings solve their respective problems in a reliable, effective, non-intrusive way and are real easy to set up and use.

Tools

Text Editor

Everybody uses one. You cannot write code without one. Whereas rich text like you might encounter in a Word document has formatting tied to it, like bolding, colors and such, plain text is just that: text and nothing more.

Notepad comes standard with Windows. It is the plainest of plaintext editors. It’s a great pick for writing Hello World, but you won’t want to be using Notepad for much longer than that.

More powerful editors like Notepad++, TextMate or gedit offer syntax highlighting, automatic formatting, and other goodies that will make your life easier. Go one step further and you’ll find editors like SublimeText and Atom, extremely capable tools that let you do things like adding text in several places at a time (multi-cursor), preview HTML or Markdown, or upload snippets to code-sharing websites. They also include features more commonly found in IDEs. This comes at the expense of startup time and memory usage, so if your specs are low you may want to start with a lighter option.

vim and emacs are very old editors with a friendly rivalry. They run in the terminal, which renders mouse use optional and makes it convenient to edit files on remote machines. They’re incredibly versatile but come with a steep learning curve.

Web Browser

Everybody uses one. Except maybe my grandma, because she’s just that hardcore. As more and more of society moves into “the cloud”, browsing the web is becoming an increasingly large proportion of how we interact with the world around us.

Web browsers are powerful bits of software. They are capable of some interesting things that help you save time and stay focused, if you know how to use them. A favorite of mine, pioneered by Opera (if I’m not mistaken) is the ability to set keyword searches. I can type w visual studio into the location bar and instantly pull up the Wikipedia page for Visual Studio. I use Wikipedia a ton, so shaving five seconds off of every wiki search I conduct adds up pretty quickly!

Take the time to learn the ins and outs of your browser of choice, and feel more like Tony Stark and less like Alice in Wonderland on the web.

Operating System

Everybody uses one. The main three are, of course, Windows, Mac and Linux (with its hundreds of varieties). Others you might not have heard of include Solaris and Haiku (formerly BeOS).

Entire books could be (have been) written about OSes. It’s probably impossible to truly master one (let alone multiple), as complex as modern OSes have become. Still, the more you know the more effective your computing will become, programmer or not. So go ahead and be a power user. Hack on your registry. Symlink your folders. Make your caps lock key do something less offensive. Just remember to go back to your real work at some point.

You may (read: will) come across the term “Unix”, which is usually used as an umbrella term for many different OSes that behave a certain way, particularly in the sense of having a sh-compatible Shell, a certain set of built-in commands, and a file system that starts with a forward slash. Mac (as a type of BSD) and Linux (all of them) are flavors of Unix. Sort of. That’s not 100% accurate, but for all intents and purposes, you can consider Mac and Linux to be cousins, after a fashion.

Communications

These are the tools you use to keep in touch with friends, family and colleagues. They include email, chat, telephony and various social media clients. Communication in the information age can be an enormous drain on your time and attention if not managed carefully.

The approach I personally take is to avoid “kitchen sink” tools like Facebook or GMail—these products combine email, chat, and social channels in a rather distracting way, and I wind up looking at cat pictures again when all I really needed to do was send a quick message. Instead I use Thunderbird for all my email, Pidgin for all my chat accounts3 and set aside time for checking social media when I can wander aimlessly and not feel bad.

I still haven’t found a decent tool for managing telephony/VoIP. Cell phones are a notch above tin-can-and-string in voice quality, and most desktop tools involve services where you pay to connect to landlines. I’d spring for a landline if I didn’t move so often. Voice is my least preferred mode of communication though, so finding a solution here isn’t a huge priority.

Productivity

Your toolbelt for Getting Things Done®. Productivity tools help you keep the gears spinning and minimize wasted time. They need not be software, either: todo lists, calendars, and contacts—the three pillars of productivity—have been around in paper form long before the personal computer was a twinkle in Turing’s eye—and there’s something to be said for no batteries required, too.

These days we’ve got Trello, Workflowy, Wunderlist, Evernote, the list goes on and on and on (pun intended). And if you can’t find a tool compatible with your brain, it’s never been easier to make your own.

Shell

Your interactive command line. At its most basic, a shell is a program that reads commands and invokes other programs.

Usually when people talk about a “shell”, they mean a Unix-style shell that follows the conventions pioneered by sh, the Bourne Shell. bash is probably the most popular of these with the more feature-packed zsh a close second.

Microsoft’s Powershell is good in its own right, but Windows-only with a high learning curve. It’s head-and-shoulders above cmd, though, so if you are invested in (or stuck with) Windows, you should get comfortable with it. Even though Windows generally favors a graphical, mouse-driven experience, being able to script things from time to time still comes in handy.

If you do find yourself on a Mac or Linux system, try fish—the Friendly Interactive SHell. It’s a more modern take on the classic Unix shell which is less compatible with scripts using “traditional” syntax but offers a more pleasant interactive experience.

Version Control System

Also known as SCCS (source code control system) or SCM (source code management), your VCS has your back when your code breaks and you can’t remember what changed. A well-wrangled VCS is almost like time travel.

VCSes usually fall under one of two categories: centralized (Subversion, Perforce) and distributed (git, Mercurial, darcs). Each has its strengths and weaknesses, but you will likely come into contact with both at some time. Notably, centralized systems differentiate between client and server, requiring you to set up a server (or access an existing one) in order to use the system.

Ten years ago, when centralized systems were the norm, it was considered good form to use one, but it was still common to go without, making changes on the fly and perhaps making copies of folders in order to back up valuable work. Today, it’s generally agreed upon that everyone should use some form of source control, and it’s extremely easy to do so—git init is all you need to get started.

Most VCSes have some concept of a repository (where the code lives), a commit (a discrete change in which you are confident and wish to “lock in”), a revert (rewinding to a previous version, either because you broke something or to gather historical information). What they are called and how they are handled varies from system to system.

A good VCS makes it easy to branch and merge your code, easing the burden of keeping track of several different changes (alternate realities?) at the same time.

git is easily the most popular VCS today, largely due to GitHub, the free service that uses git to manage (and publish, track, or discuss) code. The GitHub competitor BitBucket (run by tech giant Atlassian) uses Mercurial (hg), which is reputed to be superior to git, but GitHub’s huge lead in community adoption has made git the go-to VCS for modern development.

Build Automation

“Building” code means turning it into a runnable program you can distribute to users. In practice this can mean many different things. Compilation is the process of changing source code into binary—specifically, machine language (or, in the case of managed languages, bytecode). Minification means removing whitespace to reduce file size without changing it from plaintext. Transpilation changes source code from one language into another.

Build tools tend to be language-specific for some reason. We have CMake for C, Rake for Ruby, Cargo for Rust, and so on. The one that isn’t is also one of the oldest: make. Though it’s not specific to any one language, it does use native shell commands, so using it on Windows (while possible) can have mixed results.

You can also use a build service, or if you’re the buzzwordy type, a continuous integration (CI) thingamajangus like Travis or Appveyor, which automatically builds stuff every time you check it in to your VCS. (You are using one, right? Right?? Right.)

Deployment

After you build your code, you’ll probably want to deploy it. This might be as simple as zipping it up and uploading somewhere, or as complex as requesting computing resources, provisioning VMs, installing dependencies remotely, running deployment validation tests, and sending an email reporting how long it took to do all this…with no human intervention whatsoever.

Most deploy processes will be somewhere in between, but consider this: the former process might take a human being five minutes but a shell script written in the same amount of time can do it in seconds (with tar and scp for example); the latter process can take months if done manually, or a few hours if automated properly (with some recurring human cost to maintaining the automation setup).

My strategy is to deploy manually at first, then automate gradually those steps that turn out to be repeated.

In buzzword parlance, the deployment-focused counterpart to continuous integration is continuous delivery (CD).

Monitoring

Monitoring tools allow you to keep an eye on running software from afar and raise the alarm if anything goes wrong. I won’t go into much further detail than that because (a) monitoring software is of limited use to beginners; (b) there are so many different approaches, with many organizations rolling their own; (c) I’m not really well-versed in the arena.

Off the top of my head, well-known monitoring tech includes Splunk, Grafana, and JMX. You should ask someone else if you have an interest in the subject, because that’s really all I know.

Integrated Development Environment

Some developers prefer to use an IDE like Eclipse, Visual Studio or IDEA for the bulk of their work. An IDE combines text editing, build, and other bits necessary to make software. While this doesn’t remove the need to learn a text editor, shell, and so on, it lessens the burden by offering a working out-of-the-box setup.

By tying everything together in a known configuration, IDEs make the dev process more seamless. For example, by integrating a debugger with the inbuilt text editor (something that is possible, but difficult to achieve with standalone tools), an IDE can make following the flow of code through a complex project and pinpointing the exact point of failure into a quick and painless process.

In addition, most IDEs offer advanced features for navigating large codebases (e.g. “find references” or “call hierarchy”), intelligent refactoring (“extract method”) and code generation (declaring local variables, creating methods that haven’t been defined, or encapsulating fields via accessors). This frees the developer from busy work so they can focus on the bigger picture.

The two main downsides to using an IDE are resource consumption (all that power has to come from somewhere) and sometimes inscrutable error messages (IDEs infer a lot of information relating to the build process and often hide it from the user; sometimes they get it wrong).

The choice to use an IDE can be a controversial topic. It’s sort of a lifestyle choice. In my opinion, it makes for a poor learning environment because of all the magic that “just works” until it doesn’t, leaving the learner with no recourse. However for large, complex projects using managed languages, a solid IDE becomes indispensable for cutting through the chaff.

Miscellaneous Tools

Obviously, the space of computing tools is a vast and growing one. The ones I’ve described here barely scratch the surface of what’s out there. There’s a few key tools that I can’t resist mentioning, even though they’re somewhat specialized and don’t fit cleanly into any other category.

Vagrant is a sweet utility for managing virtual machines via Ruby. It makes it trivial to spin up a virtual Ubuntu box, and relatively easy to automate the rest of the drudge work that typically comes with wrangling VMs. Once you’re comfortable using a virtualization provider like VirtualBox or VMWare, give Vagrant a look.

Synergy is a unique client/server app that lets multiple networked boxen share a single mouse and keyboard. You simply tell it how your screens are arranged, and it detects when your mouse leaves one screen and enters another and switches control accordingly.

Finally, if you spend a lot of time in a terminal (or ever need to SSH into remote boxen for any reason) you’ll want to get familiar with a terminal multiplexer like tmux or screen. It’ll let you put terminal inside of terminals and keep them alive even if you’re not around. That might not sound very exciting, but trust me, it’s the handy little program you never knew you needed.

Soft Skills

These are skills that aren’t specific to programming or to any technical field. Entire books have been written about them. They’re among the most valuable skills to develop (since they apply literally everywhere), but are also the most difficult to study; so much so that many (e.g. creativity) one could argue more closely resemble innate talents.

I’m not sure this section truly belongs here, but erring on the side of completeness, here are a few examples of traits you should at least be cognizant of as your journey of code proceeds.

Creativity

Talent hits a target no one else can hit. Genius hits a target no one else can see.

-Arthur Schopenhauer

Don’t be satisfied to do a great job—do a great job that has never been done before, because it’s so crazy no one even considered it.

Communication

If you could be here around nine, that’d be great.

-Bill Lumbergh

Working alone is easy. Working in a team is not. Communicating clearly keeps everybody working in harmony and catches potential problems early.

Prioritization

Put first things first.

-Stephen R. Covey

Yeah, it’s tempting to go hack on that sweet particle system, but maybe you should nail down the cause of that horrific crash first so you aren’t constantly rebooting your machine in the middle of important work.

Judgment

With great power comes great responsibility.

Judgment is hard. Some of it comes from the school of hard knocks, some from the spidey-sense that certain people just have a knack for. Either way, making the right decision, the first time, without succumbing to analysis paralysis, will save you hours, money, and heartache.

Discipline

The sense of dignity grows with the ability to say no to oneself.

-Abraham Joshua Heschel

Humans are incredible reasoning machines, but they’re also scary good at rationalizing harmful behaviours and bad habits, even while acknowledging them as such. Cyclic caffeine addict? Compulsive procrastinator? Know when to put your foot down—and don’t be discouraged if it takes practice to actually do it!

Motivation

Build your trigger.

-Josh Waitzkin

Sometimes ya just don’t wanna get out of bed. And sometimes that’s okay. But most of the time it’s not. Ride the wave when it’s there. Make your own when it’s not. Whether it be music, exercise, Red Bull, whatever puts you in the zone—get your ass in gear!

Time management

Never do today what you can put off until tomorrow.

-Aaron Burr, Jr.

Time is a precious thing. 80 years is all most of us get. How you choose to spend it is up to you. Make it count.

Self diagnostics

What reason do thinkers have for thinking they think? Is it simply the fact that they think?

-Fred Dretske

My personal favorite soft skill. It’s an incredibly powerful thing to introspect, identify one’s own faults (and strengths!) and change accordingly. This sometimes comes at the expense of a small hit to the ego, but then, you have plenty of ego to spare, because you’re just that amazing. Right? Right? Of course you are.

Where to go from Here?

If you’ve read this far, congratulations, you have a longer attention span than I do. Know that whichever path you go down, you’ll have fun. My advice—take it for what it’s worth—is the following:

Google around. See what’s out there. Find software that vibes with you.

Pick a language. My preference is C, but yours may not be.

Pick a text editor. My preference is vim, but yours may not be.

Write Hello World. Run it.

Pick a piece of OPC (other people’s code). Read it. Build it. Run it. Make it do something else.

Once you feel you have mastered your first programming language and are no longer improving, choose another language in a different category. For example, if you started with C, try Lua (these two languages play particularly nicely together). Or if you started with Python, try Forth.

Do the above also with concepts, and libraries, and especially tools, rotating through the different categories as you build your knowledge and your own personal “utility belt”. It’s not a wise investment jump ship to vim if you’re already happy in another editor; experiment with your first build system instead! The law of diminishing returns and pareto principle guarantee your time will be well-spent by diversifying your learning, using a breadth of experience to guide where you choose to focus on depth.

Try to pay some mind to which technologies will complement each other, but don’t stress about it either: the real objective is that, by gaining experience in unfamiliar areas, you begin to see the knowledge you already have in a different light, and it becomes more valuable to you, simply by virtue of reinterpretation. Messing around with functional programming principles, even if you never write a single useful line of Haskell, will enable you to approach your more run-of-the-mill problems, be they at work, school, or play, with a more open mind and a stronger palette of solutions.

I think that’s pretty freaking cool.


1 If you are new to coding and find this to be utterly wrong, please let me know why! It’s been a while since I was in your shoes. I like to think I remember what it was like, but you can never quite be sure.

2 I would never recommend JS to a beginner. The language is laden with more traps than a Mayan temple and it trips up seasoned professionals on a daily basis.

3 I use AIM a lot, still. Nobody else does. I think the reason rhymes with “schmasebook”.

4 This is a poor analogy but it’s the best I could come up with. I bet you can think of a better one.

5 Modern 64 bit processors operate (mostly) on 8-byte words, but the idea is the same. You can pretend this imaginary processor is from the 8-bit era.

6 C++, confusingly, has both pointers and references, which behave almost the same way, but not quite.