Messaging Apps And Revenge Of The Computer Science Nerd

Early websites were simple HTML. They followed a pattern of a long-form equation, with perfectly balanced sides. There was no CSS to make the HTML code appear a three-dimensional hyperbole, which it is today. I hated CSS then, and I still do today.

I feel that designers, whom I both love and hate, took over my web. They took over my simple long-form HTML code and poured it in to the boiling lava of CSS and JavaScript. Outside of the web server, a lot of friends who are designers also became web and app developers.

However, as a computer science engineer, I never took on the lowly art of web design. Instead, I paid money, and lots of it, to web designers.

Web front-end development today is gobbledygook. There are 10, if not more, different ways of doing a simple thing such as rendering a form field. And many of them are indeterminate — that is, you can do a little bit with HTML, more with CSS and bring in JavaScript to screw the page 100 percent. Needless to say, the problem gets compounded when multiple JavaScript frameworks compete to manage the same markup code. God forbid, if multiple developers touch that page, it soon becomes a patient operated on by surgeons during a lunch break.

When mobile apps happened, the dependency on designers grew tenfold. Managing different screen sizes, resolutions, buttons, images and interweaving text: The overarching desire of a mobile app to look good took away the pleasure of revealing the inner beauty of bits flowing from the lake of a relational database.

A nirvana is approaching. The chat window is becoming the new user interface. Converse with an agent — human and/or machine — and get the job done. Unlike a heavyweight webpage of today, messaging apps are devoid of markup and pass raw text on the wire. HumanComputer Interaction (HCI) theory models a real-world behavior; messaging apps are helping to achieve that vision.

There’s momentum around messagingapps such as Magic, GoButler andOperator, to name a few. Many are imported from Asia, after the success of WeChat as a portal to get things done using a simple text message. From food ordering to hailing a taxi to booking a flight, many of these apps have a real human acting as a concierge, taking requests and fulfilling tasks. However, others are pure human to machine conversation. The user interaction of a food-ordering messaging app is modeled on a real-world conversation. A human talks to an agent. A machine or a human on the other side discusses menu options and completes a purchase transaction for delivery.

Chat is where computer science is shining again. The bulk of innovation and heavy lifting is happening via machine learning and data structuring rather than figuring out how to position a few elements on a web page or an app. Simple text conversation is being broken down into structured data and marshaled to an existing API via a JSON payload.

Today, most of the apps are supported by humans, sometimes an army of them. However, a lot of these tasks can be fully automated as advancements in machine learning makes technology affordable. From IBM Watson to a totally new crop of machine learning startupssuch as wit.ai (acquired by Facebook), MonkeyLearn and many others reportedly at YC’s current batch, they are building artificial intelligence-as-a-service.

With machine learning technology as a service and messaging as a user interface, now is the time again for computer science nerds. With messaging apps, we are going to get rid of the frivolity of frameworks that web pages has become, and bring back the simplicity of human-to-machine interaction.