Death of the Full-Stack Developer: Just Code Happily
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
One of my co-authors of Teach Yourself AngularJS for .NET Developers in 24 Hours asked me the other day, "Is software development actually hard?" This thought-provoking question spurred a lot of discussion. From a mental agility standpoint, development is certainly somewhere above manual labor and below rocket surgery. But can anyone and everyone learn to code? It's hard to say for sure, but numerous studies and articles have been written about the ages at which kids can learn various basic programming concepts. Others show that some people can't do it at all; certain programming concepts appear to be watershed indicators of a new developer's success in the field.
Basic concepts like variable assignment, looping, and recursion are a first major hurdle that trips up a lot of would-be developers. The second giant hurdle is the overwhelming amount of information we feed new developers. If you're a fresh-faced developer, recently graduated from college or a dev bootcamp, your work is cut out for you. On top of learning the domain of your new workplace, all the new people, and the customs and norms ("Uh, where's the bathroom?"), you're also immediately bombarded with a perpetually expanding list of software topics, languages, paradigms, options, APIs, and so on. Are we setting up new developers to be miserable?
Too Many Languages, Not Enough Time
A few months after I joined NextTier Education, we realized that we were using more languages for our core applications than we had developers on the foundational team. The front end of our web apps is written in JavaScript with AngularJS. The main API is written in Python with Django. The public-facing mobile apps are written in Objective-C (for iOS) and Java (for Android). A random back-end process is written in Go (not sure how that one snuck in there). That's five languages, with only four developers on this particular team! Fortunately, since then, we've added more developers, so we're not outnumbered.
Technical startups tend to strive toward being lean organizations. As we build each new system, NextTier chooses the appropriate technology for that system based on the information we have at that time. So while the diversity of our programming languages presents interesting challenges that potentially could be alleviated by a more streamlined tech stack, the diversity in and of itself is not a deterrent to success. The breadth of technical experience on this subset of our dev team ranges from many, many years all the way to a couple of developers who are recent graduates of dev bootcamps.
With this background, my opinions on this issue aren't catered toward teams or entire companies with the experience to handle such a diverse set of languages. Instead, my advice is meant for individual developers who might feel completely overwhelmed at the sheer pace at which our industry moves.
A Mile Wide and an Inch Deep
Is it possible to be an expert in so many languages? Maybe. All at the same time? I really doubt it. You could gradually rotate the languages in, allowing yourself to become "a mile wide and an inch deep" in all the various tech stacks. In fact, a prevailing thought among many developers is that you're not actually a good developer if you aren't a polyglot. With this mindset, in order to be a real developer you need a hat that says "Full Stack" and a T-shirt that says "Developer."
Personally, I don't think a good developer has to be a mile wide and an inch deep. Knowing five or even ten languages at once doesn't make you a "rock star" dev by itself. Good software developers write good, maintainable, scalable, and extensible code—regardless of the language or where the code lives. If you're able to accomplish that goal in Ruby, that's great. Python? Awesome. PHP? I might question your life decisions, but that's amazing, too!
But do you have to be an expert in all of those languages at once? I don't think so. In fact, I don't think it's even feasible. As Kathy Sierra says, "You're not keeping up. I'm not keeping up. And neither is anyone else. At least not in everything."
Is software dev hard? Keeping up with it certainly is. By the time I'd really gotten comfortable with AngularJS, Angular 2 was on the horizon—which is okay, because AngularJS will be around a long time, and it will remain relevant. I'd love to try out more ReactJS, learn the ins and outs of ES6 and TypeScript, start a project with Ionic, and check out MeteorJS. I wonder: By the time I get to the end of my list, will all those technologies still be relevant? And that's just in one language!
When you're faced with a daunting list of things to learn—like five programming languages—you can go about it in multiple ways.
- Grind through and learn all five languages. You might not know a lot about all five, but you can probably get your job done with a lot of Googling and reference books.
- Learn a language only as you need it. Just-in-time learning is a great way to approach a plethora of languages and frameworks. Why learn every piece of everything if you're not going to use it? Instead, find the best technology that solves your current problem, learn it, solve your problem, profit. If the next application you build needs a different language, learn that one.
- Focus on either the front end or back end. It's okay to be "just" a front-end developer or back-end developer. No matter what anyone else says, it doesn't make you less of a developer. I'll write you a note—it'll be fine.
- Focus on what keeps you interested and engaged. My favorite approach to this issue is just to learn what makes you passionate and do what makes you happy. That's when you're most productive. Some developers are perfectly happy traversing the full tech stack. Some of them are even good at it! Other developers are happier working with just a subset of technologies. If it makes you happy to do pointer arithmetic, I might start with a little prayer to the software dev gods, but I'll also say, "Good for you!"
We all need our ten thousand hours of practice, but if you're trying to learn multiple languages just so you can be a real boy, you might run out of time. Even if you spend ten thousand hours on general software development, though, I don't think it matters. For me, the "right" or "best" way to get good at anything is to love what you're doing. If you love Scheme, grab your parentheses and be the best Scheme developer you can be. No matter what language you choose, if you code a lot, you will get better.
Do I think you should barricade yourself in a dark room, focusing on only one language, particularly if you're a new developer? Not at all. Just as experiencing cultures outside of your hometown can teach you how other people live and work and what kind of food they eat, you can learn from seeing what other languages have to offer. Ooh, look, JavaScript scope isn't determined by brackets! And what are those C++ pointers even doing?! Oh, no, it's more PHP! Hide! You can learn a lot about the programming language you love, just by learning other languages. As in all other walks of life, diversity in your knowledge can only help you. You just shouldn't expect to be an expert in all the cultures of the world.
KISS Your Language(s)
So what should you do if you join a team that develops with five languages? Do you really need to become an expert in all of those languages? I think we've decided that's not essential. In fact, I'll go a step further: You don't need to be an expert in all five languages, and in some cases it makes more sense to trim from five languages to just one. To illustrate my point, let's take the five languages from that NextTier team and choose just one language that can handle all of these tasks: JavaScript.
Half a decade ago, that proposition would've been laughable, but it's much less so now. JavaScript is ubiquitous, easy to learn (and difficult to master), the community is burgeoning, and the language standard is growing up.
So let's start the teardown of the five languages.
- The front end? Already in JavaScript—done.
- The API and back-end job? You can choose from a few different node.js frameworks, and a lot of companies, giant and small, are powering their APIs with node. We're not ready to switch just yet at NextTier, but it's not outside the realm of possibility. There's a lot of concern in the industry that JavaScript isn't ready for enterprise-level application development. Many organizations have overcome that concern already, but if you're not convinced, check out some of the new things happening in the JavaScript world that should appease both type-safety fans and classical OOP geeks.
- Mobile apps? If I haven't already, this could be where I lose some of you. There have been far too many debates about HTML5 versus native mobile apps. I used to read many Twitter fights and blog-post comment wars with fascination, and even thought about taking part in a few (I generally abstained). But I'm no longer interested in such debates. Is it possible to write mobile apps with JavaScript? Absolutely. Is it always the right choice? Nope. For our purposes, NextTier could most likely leverage something like Ionic to get the results we want. I've been on a team that wrote mobile applications generating millions of dollars in revenue, and it was 98% JavaScript. If we need to write something that absolutely demands performance, like a mobile game, JavaScript probably isn't the answer. As the language is growing up, though, more often than not JavaScript is becoming the answer.
One Language to Rule Them All?
At this point, I've outlined how we can drop four languages and ride the JavaScript train to the bank. I'm not advocating that you march into your CTO's office and go on a hunger strike until you're using JavaScript for everything. It's okay—and most likely even appropriate—for your company to have language diversity. I'm simply illustrating the point that as an individual developer, it's not only possible to get by without comprehensive knowledge of half a dozen languages, their frameworks, and their APIs—actually it's quite feasible.
Some of you aren't convinced. Let's back up again to our five programming languages. You would likely be able to get some knowledge overlap from your team members on a few of these projects, but across all five languages? That's really unlikely.
If all the projects used JavaScript, though? The likelihood goes way up. Yes, writing an API is a different paradigm from writing a web front-end in Angular, which is different from writing a mobile app. No doubt about that. But if you're using the same language, concepts will overlap. Even if you hate the idea of a one-size-fits-all language, you have to be intrigued by the idea of gaining efficiency by cutting your languages from five to one.
But what if JavaScript isn't for you? That's okay. You could also reduce your five languages to one or two with C#.NET and Xamarin. I'm sure other solutions would also work.
Choices are good, because no language is a magic bullet. Sometimes you need to eke out the best possible performance from a system, or solve a very specific problem. I'd never suggest JavaScript for statistical analysis—R is a great solution for that. For AI programming, Visual Basic isn't as good a tool as LISP. Programming languages are tools to solve problems. Be wary of thinking of every problem as a nail, just because all you have is a hammer.
Let's be clear: You should learn to use other tools in your tool belt. But if you don't love doing it, solve the problem that requires the tool, and get back to what you enjoy doing.
Final Thoughts
There's no great software developer award for knowing the most languages or technologies or frameworks—or at least I didn't get invited to the ceremony. If the idea of five languages genuinely excites you, and you want to learn all five because it sounds like fun? Outstanding! If you're consistently able to learn and maintain your skills with a bevy of languages, while being amazing on the front end and back end, I'll admire you forever. I don't think I can pull that off, though. Instead of being pretty decent at a lot of languages, I'd rather aspire to be excellent at just one.
