How Peach’s Most Interesting Feature, the Hybrid Command Line, Is Becoming Mainstream Again

Photo: Adrianna Williams/© Corbis. All Rights Reserved.

Shortly before the end of the workday last Friday, an app called Peach suddenly became the hottest social network among a small but influential coterie of digital media industry watchers. It’s tough to say precisely what Peach is. It’s a social network that lets users post all types of media, but feeds aren’t public, nor does it have an aggregate feed similar to a Facebook News Feed — every user is siloed.

Peach is fun to play around with, but mostly what it has going for it as a social network is that it’s not Twitter or Facebook. The one interesting thing about Peach is a feature that its makers — which include a co-founder of Vine — call “magic words.” Magic words are shortcuts that allow users to access new types of posting features. Typing “GIF,” for instance, lets users search for the appropriate media to post; “battery” offers a way for users to post their current charge percentage; “move” lets users access their phone’s motion tracking to post the number of steps or distance traveled.

Magic words” are easy, and they feel fresh and new compared to the button- and menu-heavy interface for posting status updates to Facebook. But they’re about as old a user interface as you can imagine. Peach is the latest app to turn away from a pure GUI, a “graphical user interface” that relies on buttons and icons, and back toward a watered-down version of a CLI, the text-based “command line interface” that was the defining method of using personal computers until the 1980s, joining much bigger peers like Apple, Microsoft, Slack, and, yes, Facebook in an attempt to rethink how we interact with our various devices by looking at computing’s past.

The early days of computing (at least, computing that required a screen) used the command line as the singular method of user input. Users typed their commands as text. It wasn’t until Apple began manufacturing computers — with a clickable, menu-based interface famously stolen from Xerox — that alternative, graphical systems came to prominence. The command line argument is still used today by programmers, but since the mid-1980s, the GUI has been the dominant mode of consumer-facing personal computing — in particular, a subset of GUI known as WIMP (windows, icons, menus, pointer).

There are obvious advantages to the WIMP system. It makes understanding how a computer works simpler by attaching physical-world metaphors to all actions. We’ve gotten so used to WIMP systems that it can take a minute to remember the extent to which its basic building blocks are formed of analog analogues: “Files” are placed in “folders,” “deleted” items go in the “trash,” actions are compartmentalized into taskbar “menus” at the top of the window, and icons like the floppy disk reflect actions such as saving a file (maybe it’s time to update that one).

Using a computer with a GUI, which requires lots of moving and clicking, tends to be somewhat slower than using one with a command line, but the learning curve is far steeper on a command line interface, where visual metaphor doesn’t exist and every action requires knowing the specific command. Clicking through, copying, and creating folders on your PC requires a handful of clicks or keyboard shortcuts. In the command line, it requires knowing commands like “cd,” “cp,” and “mkdir.” On a WIMP operating system, the help menu is ever-present at the top of the screen. In a Unix shell, typing the “help” argument pulls up this:


Mastering the command line has clear benefits. Many coders still swear by it, claiming that the WIMP system only slows them down. Working within a GUI, for instance, uses up more of a computer’s hardware than a text-based interface. As Wired wrote in 2012, “geeks like command lines just because you have to know what you’re doing to use it. You have to know the commands. You can’t hunt and peck like you do with a GUI.” This idea of mastering the command line can be seen in other industries besides software and hardware development. The Bloomberg Terminal, widely used within the finance industry, has an atrocious appearance, and some power users — prideful of having mastered the inelegant system — navigate solely via keystroke.

But in recent years, apps have increasingly found ways to combine the command line interface with the GUI text box, creating a hybrid that merges the GUI’s approachability with the CLI’s power. Peach’s magic words are a good, simple example: A few years ago, you might have seen every action tied to a magic word included in a menu interface, letting you scroll through and peruse all of your options. Peach does away with this, though — just start typing, and if the text matches up with a command, the app will let you perform it.

Peach is far from the only app to move toward a more free-form approach to human-computer interaction. In Mac OS X, hitting ⌘-space brings up Spotlight Search, which has broadened its scope in each iteration. Originally meant to search local files on your hard drive, Spotlight can now parse text inputs to launch apps quickly, search for movie showtimes, look up information on Wikipedia, run mathematical calculations and conversions, etc.

The popular chatroom/collaboration software Slack uses slash commands similar to what its spiritual predecessor IRC used. Typing “/away” into the text field doesn’t send that message to your co-workers, it just toggles your status. These commands are extendable, and people have made all kinds of custom slash commands and slackbots, the idea being that it keeps user interaction focused around the text field.

The text field is key. It’s partly the solution to a problem that has long plagued mobile developers: Less screen real estate means less space to present interaction options to the users. But it’s also that, for the first time, engineers and and developers are creating products for a population that’s truly digital native — for whom typing comes naturally and for whom digital actions don’t need to be metaphorized.

Now that the computer most people use most often is carried in their pockets and doesn’t have a “desktop,” WIMP’s strengths are less important. And made a little more friendly looking and stripped of its connotations as a nerd’s power tool, text input is an even lower barrier to entry than any GUI: The only limit is your vocabulary. Now that computational power is available at a lower price than ever before, it’s possible to create pseudo-command-line systems with greater leeway for mistakes, and even the limited ability to “learn.” (In other words, you don’t need to learn a whole manual’s worth of commands to use it effectively.)

In the command line, you had to have your options memorized; in the GUI, you could browse the options. But in a new space of user experience, there is no clear set of options for interacting with software. There is, if software continues to track in this direction, no command that a computer won’t be able to process, and that will make them subsequently easier to use, and more vital than ever.

The experience is so important to some that for initiatives like Facebook’s M service, artificial intelligence is supplemented by the work of actual human beings on the other end. And this is the place where Facebook might feel a little threatened by Peach. While we tend to think of Facebook as a social network defined by its News Feed, the company is pouring more and more resources into its stand-alone messaging app, turning it and “M” into a way to accomplish tasks like ordering an Uber. A social network is worth a lot of money. But a service that can insert itself into a middleman in nearly any transaction is potentially worth even more.

And this, ultimately, is the future of user interface. We’re coming out of a 30-year period in which personal computers obeyed the rules of a particular metaphor, the office desk, and into one in which they’re going to obey another, the personal assistant. That includes “digital” assistants like M, Apple’s Siri, Microsoft’s Cortana, and Google. Advanced voice recognition and intelligent text parsing don’t just get a user relevant information or perform an action faster than using a GUI might — they can also become a new layer of services from which money can be extracted. Peach isn’t there yet, and probably never will be. But if I were a venture capitalist I wouldn’t be asking them about how they plan to compete with News Feed — I’d be asking them how they plan to compete with M.