Archive for August, 2010

How to Become a JavaScript Badass

August 16th, 2010 by Aaron N.

Last week I wrote a rather lengthy post lamenting the fact that finding talented front end developers is rather challenging. This led to some nice comments and emails about the topic and more than one suggestion that I not just whine about the state of things without suggesting a remedy; specifically, that I post instructions on HOW to become a JavaScript badass.

I don’t know that I can tell you how to become a JavaScript ninja, but I can give you some pointers. I’ll start out by saying that I learned what I know via a rather circuitous route that I doubt even I could reproduce if I tried…

My Story, In Brief

My road to being a JS developer is not that interesting, if you ask me, but, shrug if you want to know, here are the highlights:

  1. I studied music, photography, and graphic arts in high school and some in college, though it wasn’t my major (which was Jazz Guitar – seriously).
  2. I learned HTML back when Mosaic was still in use and grew my web skills as they arrived in each new browser update.
  3. I joined a startup (knowing HTML in 1997 was a hot commodity) where I learned a bit of programming (Java). I wasn’t really very good at it. I built a really simple CMS with it.
  4. In 1999, I somewhat accidentally started a music startup (http://www.epitonic.com) which I designed and built. I built a much more robust CMS w/ a very small dev team. LOTS of vanilla JavaScript.
  5. In 2003 I joined CNET as a product manager. I didn’t write much code, but whenever we needed some JavaScript, I rolled up my sleeves and pitched in. I inadvertently became “the JavaScript guy” there and started blogging about my research on clientside.cnet.com.
  6. At first focused my work around Prototype.js and Moo.fx.
  7. When MooTools first arrived, I was hooked from the start (here’s posts one and two from 2006 of my first thoughts on the framework).
  8. I got involved with the project to help drive adoption at CNET. I became a committer and then put together classes to teach developers there (at CNET). This further cemented my role there as “the JavaScript guy” though I still thought of myself (and still do) as a product manager.
  9. I left CNET in late 2007 to start Iminta.com, which I’m still proud of. Then the economy died at the end of 2008 and I started contemplating finding a job and found Cloudera (or, really, they found me).

I know, that’s not terribly interesting or informative, but it gives you an idea of what a bizarre roller coaster my professional career has been. For me JavaScript and web development in general has always been a means to an end. I like building interesting products. Writing some JavaScript or some server-side code has always been about making a good web experience but I don’t really care if I write the code or not. I like to write code, but it’s not my objective. My job at Cloudera is the first job where I have principally been an engineer, though you could argue at my two startups that was my job, I’d retort that my job was to design the product and it just so happens that I helped build it, too.

Enough About Me, Let’s Talk About Becoming a JavaScript Ninja

All this about me is really just to put into context the fact that I can’t tell you the right way to become a front end developer. You will have some doors open in front of you as you pursue your goals and they won’t be like the doors opened infront of me. CNET basically paid me to spend roughly a year studying JavaScript and JavaScript frameworks in detail, and that’s a hard job to find. That said, I do think you can reproduce a lot of it on your own and, in some ways, I learned a lot of this stuff the hard way. Without any further equivocating, let me tell you the kinds of things I think you can do to get these skills.

  1. Study design and designers. I’m not saying you have to have the talent to be an awesome graphic designer, but you should pay attention to people who are. When you are surfing the web, pay attention to what works. What looks good? What communicates to you that you can do something on a page? A great example here is this video of Bill Scott’s work on UI patterns. He gives great examples of what the “interesting moments” are in UI design. But in general, you should pay attention to the sites you visit and notice when they get things right and wrong. I often ask interview candidates what sites they admire and why.
  2. Study JavaScript. I mean really dig into it. Watch all those awesome Crockford videos – ALL OF THEM! – on the YUI theater. For that matter, watch all the OTHER videos there. Seriously awesome stuff. I don’t agree with 100% of what they all say, but they are educational for sure. Read Crockford’s JavaScript: The Good Parts. Again, I don’t agree with it 100%, but it’s a seriously solid overview of the language.
  3. Study JavaScript Frameworks. Note that this is plural. The single most important thing I’ve done in my education with the language was to write the original documentation for MooTools. To do this, I had to read the entire library’s undocumented source and figure out what it was doing and why. I’ve learned a lot since then, but nothing I’ve done has ever resulted in as big a jump in my knowledge. When I wrote jqueryvsmootools.com I did it again, this time with jQuery. I read the entire source so that I could understand it. I did it again with Dojo when I put together a talk about programming to patterns that I first gave in tandem with Dylan Schiemann of the Dojo team (the talk wasn’t about frameworks so much as it is about the value of abstraction, and I wanted to make sure it wasn’t just a MooTools focused talk). I’ve done the same thing with other frameworks to learn the lessons that I can from other people’s development styles. Don’t just use jQuery or YUI or MooTools. You need to study all of them to understand what makes them the same, different, and interesting. Don’t stop until you understand everything that these frameworks are doing and, more importantly, why. Don’t hesitate to ask their authors for explanations; most are happy to talk about their work.
  4. Get involved with a framework. The second most important thing I’ve done with JavaScript was getting involved with the MooTools project itself. Working with open source projects is a HUGE boost to your resume and, here’s the thing, you don’t need to really know that much to get started. You just have to be willing to spend the time. Right now, there are dozens of bugs open on every framework out there. Go fix some! Go write test cases! Go write a blog post about how you use it! Do these things and get committer status and I promise you you’ll start getting a ton of interesting job offers.
  5. Release some of your own code. I can’t stress this enough. If you don’t have code on github (or google code or your own site or whatever) you’re wasting a big opportunity. Releasing your own code allows me, a potential employer, to know your capabilities before I hire you. This stuff gets people interested in you. If you release a lot of your code, you may even get others to help you maintain and grow it. This is how open source projects get going. I almost consider it a red flag to see a resume without a url to github or something similar.
  6. Blog about it. Write down everything you learn as you learn it on your blog. Next thing you know, 3 years have gone by and you have this huge body of work. Stuff on your site draws the attention of other developers who are struggling with similar problems. You become an expert without really meaning to. If you blog constantly about what you are doing, what you are studying, you’ll find that people come to you with work to be done expecting that you are awesome because, well, you’re explaining all this stuff to them. I can’t stress the value of this enough, though it is very time consuming. This is especially valuable if you’re a freelancer.
  7. Build something interesting. I once spent a month or two writing a photo gallery in PHP just to have an excuse to learn PHP better. I learned Smarty with that little project, too. I’ve built a lot of things for the excuse to learn it. I built Iminta.com with a friend and we chose Ruby on Rails mostly because neither of us had built anything with it and wanted to learn it. Forcing yourself to do things with new languages and environments will grow your skills faster than anything. Don’t rely on the skills you have; always look for chances, excuses really, to do things in new ways. Working with emerging technologies will make you debug that technology itself and maybe contribute fixes back to it. It can be painful, but it also makes you really learn how that technology works.
  8. Join a startup. I know, this one can be tricky, especially if you don’t live in the SF Bay area. But joining a startup will make you tackle problems that aren’t in your domain because, well, there’s no one else to do it. If you’re not experienced enough to be the 2nd or 3rd person at a startup, aim for being the 10th or 20th. You’ll be asking for long hours and low pay, but you’ll get a mountain of experience. Think of it as an extended college education that pays you a little cash and (if the stock takes off) might buy you a house.
  9. Take the time to learn why solutions work. When you’re working on something and you get an error and find the solution on Google, take the time to really understand what the problem was. When you are starting up an app server – ruby on rails, django, lamp, whatever – and you get a stack trace, take the time to dig into it and understand the problem and what the logs are telling you. Debugging stuff on the command line will teach you a ton. It’s slow, thankless work, but it’ll greatly improve your value when you take on more challenging tasks at new jobs.
  10. Be curious, and fight off laziness. This is a bit of a weird one, I know. What I mean by it is that you should look at tasks that require you to do new things as opportunities. Recognize when these moments come along and cherish them. There is nothing more awesome than having a job that pays you to learn. If you have coworkers that know things that you don’t, and vice versa, trade them. When I was in college I told the guy who was building the web site for my school that I’d help him design it and show him how to use Photoshop if he would teach me HTML. I joined Cloudera 18 months ago and knew zero Python and now I’m pretty decent at it. If you have a job that uses technologies you don’t know, don’t just stay in your little JavaScript world; find ways to expand your knowledge however you can.

I could probably go on about this stuff for a while longer, but I’ll stop. What it really boils down to is that you have to want to be a front end developer and pursue it the way a concert violinist pursues the first chair. There are hours and hours of practice for which no one will pay you. But eventually you’ll find yourself in a position where someone will pay you to do what you love and you’ll find it hard to believe how you got there. Next thing you know, you’re the person designing the user experience at some hot startup and trying to hire new people to join you.

My last suggestion to you, if you’re a budding JavaScript badass, is to force yourself to find a job that will help you learn these skills. Don’t just take a job making some company’s corporate website. There are tons of jobs out there that need front end talent that are building interesting things.

Did I mention Cloudera is hiring?

Posted in Deep Thoughts | 30 Comments »

Why It’s a Good Idea to Be a JavaScript Developer, and What it Takes to Be One

August 13th, 2010 by Aaron N.

So, I’ve been trying to hire another JavaScript developer at Cloudera for a while now and the effort has proved to be rather daunting. Finding really good JavaScript developers who know their stuff is hard! If anything, it’s driven a very simple point home to me: it is a good time to be a JavaScript badass. In the past few months, when I’ve found talented JS developers and tried to recruit them, it’s clear that they have as many doors open in front of them as they want. If anything, they’re bombarded with requests from recruiters and companies trying to recruit them. As a person trying to hire these people, I find myself competing in a way that reminds me of the first dot-com boom. I offered one candidate a 1967 Cadillac Eldorado and still didn’t close the deal (I wasn’t serious of course, but I did offer them a lot, including the car, but the candidate new I wasn’t serious about that one. He countered by demanding something a little more high-profile). In the end he took a job at another startup.

The thing is, being a JavaScript badass isn’t only about knowing JavaScript. If you want to be a solid front end developer, you need to be able to do a lot of things really well, most of which aren’t really in the job description or, at least, don’t appear to be as important as just knowing JavaScript. Some of them just require competency, while others really require mastery (JavaScript does, obviously, if you want to qualify as a certified badass). In general, I’m a relatively humble person. I have my own blog, which I guess says something about the size of my ego, but I’m not that full of myself, I swear. But consider the fact that when I was recently talking to a recruiter about the position, I described what we were looking for, the skills and the kind of person, and the recruiter told me flat out, “Those people don’t exist; you’re describing a unicorn.” I told the guy I was describing the job that I do every day. He was unyielding, “Well then consider yourself mythical in nature.”

So what does it take to be a JavaScript badass, and what do you get for it? Well, to be a really top-notch front end developer means you’re going to have to spend a lot of time doing stuff that’s tedious and boring, some stuff that’s really hard and beyond your talents, and some stuff that’s just plain fun. For what it’s worth, here’s my list:

  1. You need to know HTML. You need to know everything about it. Everything. You need to know what effect document types are going to have on your markup’s display. You need to know that browsers will shove a TBODY tag into your tables even if you don’t code one. You need to watch what the browser vendors are adding into their nightly builds and their upcoming specs. Learning this stuff, at least for me, just takes loads of time discovering it as you work.
  2. You need to know CSS so well you’ll regret it. These days most of the cross-browser problems I encounter are with their CSS implementations. Getting around them usually means describing the exact same style in a different way. You just added a negative margin-top to a link and guess what, the users of certain browsers can no longer click them, despite the fact that you gave it a z-index of a million and you can clearly see the damn thing. Or you just made the most amazing collection of positioning statements to make an auto-resizing masterpiece of CSS awesomeness but god help you if your boss looks at it in Internet Explorer. The thing is, having this skill – being good at CSS – isn’t really all that interesting or valuable any more. Everyone who works on the front end is just expected to have it. There are some seriously talented CSS developers out there who go far above and beyond this kind of knowledge and truly make CSS into an artform. You don’t have to be one of them, but if you can’t make the page look like the mockup, it doesn’t matter that you can do just about anything else.
  3. You need to know the browsers. You need to know that IE won’t let you convert this string: “<meta></meta>” into DOM elements by setting the innerHTML of a DIV. And forget about injecting it in the HEAD to turn it into a DOM element, because that thing is read only. You need to know that IE will always barf at trailing commas in your array and object literals. You need to know that IE has offset positioning issues for elements inside of iframes. And IE isn’t the only pain in the ass; Firefox has those clearfix issues with floated elements and webkit’s no saint. And you need to know that the latest versions of webkit have whatever awesome new features it has, as well as what ways mobile safari differs from the desktop version.
  4. Did I mention JavaScript? You need to know how ALL of JavaScript works. You need to know that the arguments object in a function has a length property and is iterable, but it isn’t an array and has no methods. Further, you need to know how to turn it into one. You need to know how the prototype property works, how to use it on your own without the help of a framework, and what the implications of using it are. You need to know how to garbage collect things and, for that matter, how to tell when you aren’t. You need to be really good at abstraction. Seriously brilliant at it. You need to be good at designing APIs for yourself as well as others. You need to know how to test your code and how to manage it.
  5. You need to be decent at design – information flow, graphic design, user experience design, code design. You need to be able to interview a user, a client, a stranger, an expert, whatever, and ask them the right questions to be able to devise an experience that solves their problems. Not just the ones you can imagine, but the ones they don’t know how to describe. You need to be able to open Illustrator and put together wireframes quickly to devise a plan, then switch to Photoshop – god you need to be good (that is, efficient) at Photoshop – and put together the visual style. You don’t have to be the worlds best designer; I’m certainly not. But you need to be able to tell the difference between good design, bad design, and amazing design. If you can manage good on your own, you’re set.
  6. You need to be comfortable on the server side. There aren’t that many positions out there where all you write is JavaScript all day long. Most front end jobs will see you writing as much as half of your code in JavaScript, but you need to be able to pick up other languages relatively quickly (in, say, a month or less). Python, Ruby, Java, PHP… If you only know one programming language, it doesn’t count. If you only know JavaScript and PHP, and you balk at any task that requires you to use something other than those two (rather than, say, looking at it as an opportunity to learn something new), then you need to ask yourself what you’re doing here. Knowing how to use a framework, but not how to accomplish something without one – with vanilla JavaScript – doesn’t count as knowing JavaScript.
  7. And then there’s about a million other things that you need. You need to understand Git. Git is amazing. I know you’ve heard people tell you this, and if it still seems like an alien monster that refuses to explain itself, then you need to hunker down and figure it out. You need to understand how to get things done with just a shell. You should be able to open up vi or emacs and edit a conf file without launching notepad or TextMate or whatever. You should be able to run make and compile things. You should be able to tail a log file, unzip a tarball, secure copy things, run servers and configure them. You need to be able to kill processes and read a stack trace.

I’m not even counting stuff like being familiar with Agile development processes, working with open source communities, or releasing code. But if you feel like out of the seven things I listed above that you nail 5 of them and would get passing grades at all 7, then congrats. You’re a unicorn.

And what do you get for being one? How can all these things possibly be worth the years of tedium, of debugging crap in IE6 and worse (yes, there is worse), possibly be worth it? Well, you’ll get recruiter emails and calls every week (this isn’t so awesome, actually; they’re like sales people who pretend they know you). When you want to look for a new job you won’t even need to do much – just tell a few people or post it on twitter – and you’ll have half a dozen doors open to you. You’ll have friends that are starting new ventures calling you, begging you to join up. You’ll get in on that ground floor you’ve been hearing about. You’ll get good stock options and you’ll get to design products from the ground up. You’ll get a 1967 Cadillac Eldorado as a signing bonus (not really). That guy I offered that car to (not really) took a job at a startup and he’s the guy. He’s THE front end developer for their product which doesn’t exist yet. He gets to point to that thing, a year or two from now when it’s worth a bazillion dollars (I do wish him luck, after all) and say, “I did that.” There are people out there, right now, who get to point at Twitter, Facebook, Gmail and Google Maps, at the Iphone’s UI, at Github, at the YouTube player – stuff used by millions upon millions of people – and say, “I did that.”

That could be you.

Did I mention Cloudera is hiring?

Posted in Deep Thoughts | 29 Comments »

My Talk at JSConf: Programming to Patterns

August 3rd, 2010 by Aaron N.

A few years ago I wrote a post after visiting the Ajax Experience about Programming to Patterns. The JSConf team was nice enough to bring me out to DC to give a talk, which you can watch below, and you can get the slides in PDF or Keynote format here.

Posted in Best Practices, Organizing Code | 4 Comments »