Programmers love to sneer at the world of fashion where trends blow through like breezes. Skirt lengths rise and fall, pigments come and go, ties get fatter, then thinner. But in the world of technology, rigor, science, math, and precision rule over fad.
That’s not to say programming is a profession devoid of trends. The difference is that programming trends are driven by greater efficiency, increased customization, and ease of use. The new technologies that deliver one or more of these eclipse the previous generation. It’s a meritocracy, not a whimsy-ocracy.
What follows is a list of what’s hot and what’s not among today’s programmers. Not everyone will agree with what’s A-listed, what’s D-listed, and what’s been left out. That’s what makes programming an endlessly fascinating profession: rapid change, passionate debate, sudden comebacks.
Not: Full language stacks
It wasn’t long ago that people who created a new programming language had to build everything that turned code into the bits fed to the silicon. Then someone figured out they could piggyback on the work that came before. Now people with a clever idea simply write a preprocessor that translates the new code into something old with a rich set of libraries and APIs.
The folks who loved dynamic typing created Groovy, a simpler version of Java without the overly insistent punctuation. There seem to be dozens of languages like Scala or Clojure that run on the JVM, but there’s only one JVM. You can run many languages on .Net’s VM. Why reinvent the wheel?
This isn’t exactly true. The hypervisors have their place, and many Docker containers run inside of operating systems running on top of hypervisors. However, Docker containers are soooo much smaller than virtual machine images, and that makes them much easier to use and deploy.
When developers can, they prefer to ship only Docker containers, thanks to the ease with which they can be juggled during deployment. Clever companies such as Joyent are figuring out how to squeeze even more fat out of the stack so that the containers can run, as they like to say, on “bare metal.”
In the digital business era, you need agility to seize new opportunities while efficiently maintaining legacy, on-premise systems.
There are now dozens of frameworks like Kendo, Sencha, jQuery Mobile, AngularJS, Ember, Backbone, Meteor JS, and many more, all ready to handle the events and content for your web apps and pages.
Those are merely the web apps. There are also a number tuned to offering cross-platform development for the smartphone/tablet world. Technologies like NativeScript, PhoneGap, and Sencha Touch are a few of the options for creating apps out of HTML5 technology.
Hot: CSS frameworks
Not: Generic Cascading Style Sheets
Once upon a time, adding a bit of pizzazz to a web page meant opening the CSS file and including a new command like
font-style:italic. Then you saved the file and went to lunch after a hard morning’s work. Now web pages are so sophisticated that it’s impossible to fill a file with such simple commands. One tweak to a color and everything goes out of whack. It’s like what they say about conspiracies and ecologies: Everything is interconnected.
That’s where CSS frameworks like SASS and its cousins Compass have found solid footing. They encourage literate, stable coding by offering programming constructs such as real variables, nesting blocks, and mix-ins. It may not sound like much newness in the programming layer, but it’s a big leap forward for the design layer.
Hot: Video tags
Not: Static tags
Once upon a time, video was something you watched on YouTube or Vimeo. It was a separate thing that lived on its own in a dedicated page. That’s changing as more and more websites use video as building blocks like static GIFs or JPGs. All of a sudden, the screen starts to move as the people or dogs come alive.
Hot: Almost big data (analysis without Hadoop)
Not: Big data (with Hadoop)
Everyone likes to feel like the Big Man on Campus, and if they aren’t, they’re looking for a campus of the appropriate size where they can stand out. It’s no surprise then that when the words “big data” started flowing through the executive suite, the suits started asking for the biggest, most powerful big data systems as if they were purchasing a yacht or a skyscraper.
The funny thing is many problems aren’t big enough to use the fanciest big data solutions. Sure, companies like Google or Yahoo track all of our web browsing; they have data files measured in petabytes or yottabytes. But most companies have data sets that can easily fit in the RAM of a basic PC. I’m writing this on a PC with 16GB of RAM—enough for a billion events with a handful of bytes. In most algorithms, the data doesn’t need to be read into memory because streaming it from an SSD is fine.
There will be instances that demand the fast response times of dozens of machines in a Hadoop cloud running in parallel, but many will do fine plugging along on a single machine without the hassles of coordination or communication.
It’s not so much that Hadoop is cooling off. It’s more that Spark is red hot, making the Hadoop model look a bit old. Spark borrows some of the best ideas of Hadoop’s approach to extracting meaning from large volumes of data and updates them with a few solid improvements that make the code run much, much faster. The biggest may be the way that Spark keeps data in fast memory instead of requiring everything be written to the distributed file system.
Of course many people are merging the two by using Spark’s processing speed on data stored in Hadoop’s distributed file system. They’re more partners than competitors.
Hot: Artificial intelligence/machine learning
Not: Big data
No one knows what the phrase “artificial intelligence” means, and that helps the marketers, especially since the term “big data” has run its course. They’re grabbing terms from artificial intelligence and upgrading the sophistication of the big, number-crunching algorithms that plow through our log files and clickstreams. By borrowing the more sophisticated algorithms from the 50-odd years of AI research, we stand a better chance than ever of finding that signal in the noise. Tools run the gamut from machine learning frameworks to cognitive computing, all the way up to IBM’s Watson, which you can now ping to solve your problems. Each offers its own level of machine intelligence and, with this, the promise of taking over more of the data analysis and forensics for us.
Not: Virtual things
Was it only a few minutes ago that we were all going to be living in virtual reality where everything was drawn on our retinas by some video card? It still might happen, but in the meantime the world of robotics is exploding. Every school has a robotics team, and every corner of the house is now open to a robotics invasion. The robot vacuum cleaners are old news and the drones are taking off.
That means programmers need to start thinking about how to write code to control the new machines. For the time being, that often means writing scripts for lightweight controllers like the Raspberry Pi, but that’s bound to change as the libraries grow more sophisticated. Many roboticists, for instance, like hacking the code in OpenCV, a machine vision platform filled with C. This means new rules, new libraries, new protocols, and plenty of other new topics to think about.
Hot: Single-page web apps
Remember when URLs pointed to web pages filled with static text and images? How simple and quaint to put all information in a network of separate web pages called a website. The design team would spend hours haggling over the site map and trying to make it easy enough to navigate.
New web apps are front ends to large databases filled with content. When the web app wants information, it pulls it from the database and pours it into the local mold. There’s no need to mark up the data with all the web extras needed to build a web page. The data layer is completely separate from the presentation and formatting layer. Here, the rise of mobile computing is another factor: a single, responsive-designed web page that work like an app — to better avoid the turmoil of the app stores.
Hot: Mobile web apps
Not: Native mobile apps
Let’s say you have a great idea for mobile content. You could rush off and write separate versions for iOS, Android, Windows 8, and maybe even BlackBerry OS or one of the others. Each requires a separate team speaking a different programming language. Then each platform’s app store exerts its own pound of flesh before the app can be delivered to the users.
Or you could build one HTML app and put it on a website to run on all the platforms. If there’s a change, you don’t need to return to the app store, begging for a quick review of a bug fix. Now that the HTML layer is getting faster and running on faster chips, this approach can compete with native apps better on even more complicated and interactive apps.