It was a sunny afternoon.

There I was: optimistic, and ready to take on the software world. I was encouraged by the vast number of from-scratch projects implemented without bloated, proprietary frameworks, by developers who weren’t afraid to understand the physical machine that they were working on.

I downloaded one of these projects to give it a whirl. I was playing around with the custom user interface and using it to build some data. Feeling happy about it, I went to save my data to disk. The user interface prompted me for a file path. I typed a filename—“ryans ptojecr”—“Oops! Made a mistake. I meant to type project. Let me hit Ctrl + Backspace and try that again.”

Immediately after this keypress, on the screen it showed me “ryans ptojec”. The user interface programmer has not implemented Ctrl + Backspace.

I became grumpy. That is the moment that I knew I had to write this post.

A Well-Earned Reputation

Custom, from-scratch user interfaces have earned the reputation for lacking a number of standardized features. This reputation is well-earned. I won’t defend all feature standardizations, but some of them exist for a reason. A great example is the set of standard keyboard shortcuts within a text field, which a substantial percentage of current computer users are familiar with. They rely on these shortcuts for quickly communicating their intent through a user interface.

Arguments I’ve heard for lacking support for such shortcuts often include a perception of the problem being very complex, and that it’d require more code than practically required.

I know from experience that this is not the case. I won’t claim that I’ve implemented literally all shortcuts—there are a lot of them—but I’ve implemented a fairly large subset of them, and the problem can be simplified such that it requires very little code.

Let’s take a look at how.

Getting Our Bearings

First, I’d like to list all of the behaviors that I’ll cover in this post. It won’t be all of them, but it’ll be most of the common ones, and the implementation I provide in this post will be easily extendable to support more.

Let’s first define some terms, so that each behavior can be well-defined:

Point — Some position within the text—exists between characters.
Cursor — A point within the text that changes directly with user input, to position new insertions, deletions, or selections.
Selection — A contiguous range of points within the text that the user actively changes in order to delete, copy, or replace chunks of text at once.
Word — Application-specific; some chunk of text that is naturally separated from other words. In a program in which you modify natural language text, this may be alphanumeric chunks of text separated by whitespace. Code editors tend to have slightly different rules—for example, they will consider whitespace, but also characters like _, /, or other symbols to also break words apart.

With that out of the way, here are all of the behaviors I’ll cover:

User Presses Right → If there is no selection active, the cursor moves one character to the right. If there is a selection active, the cursor is moved to the maximum side of the selection range.
User Presses Left → If there is no selection active, the cursor moves one character to the left. If there is a selection active, the cursor is moved to the minimum side of the selection range.
User Presses Shift + Right → The current selection—including if it’s empty—is kept, and the selection endpoint defined by the cursor position is moved to the right.
User Presses Shift + Left → The current selection—including if it’s empty—is kept, and the selection endpoint defined by the cursor position is moved to the left.
User Presses Ctrl + Right → The current selection is removed. The cursor moves to the next word beginning, where word can be defined by the application. There is not a fully standard behavior here—code editors, for example, will consider words to begin and end at a different place than your web browser address bar does.
User Presses Ctrl + Left → The current selection is removed. The cursor moves to the previous beginning of a word.
User Presses Ctrl + Shift + Right → The current selection—including if it’s empty—is kept, and the selection endpoint defined by the cursor position is moved to the next word beginning.
User Presses Ctrl + Shift + Left → The current selection—including if it’s empty—is kept, and the selection endpoint defined by the cursor position is moved to the previous word beginning.
User Presses Backspace → If there is no selection, the character to the left of the cursor is deleted. The cursor moves back one character. If there is a selection, the entire selection is deleted, and the cursor moves to the beginning of the previous selection range, and the selection is removed.
User Presses Delete → If there is no selection, the character to the right of the cursor is deleted. The cursor stays in the same position. If there is a selection, the entire selection is deleted, and the cursor moves to the beginning of the previous selection range, and the selection is removed.
User Presses Ctrl + Backspace → If there is no selection, the range from the cursor’s position to the closest word beginning to the left is deleted. The cursor is moved to the minimum position of the deleted range.
User Presses Ctrl + Delete → If there is no selection, the range from the cursor’s position to the closest word beginning to the right is deleted. The cursor is moved to the minimum position of the deleted range (in this case, it does not move).
User Presses Home → The cursor moves to the beginning of the line. The selection is removed.
User Presses End → The cursor moves to the end of the line. The selection is removed.
User Presses Shift + Home → The cursor moves to the beginning of the line. A selection is formed from the cursor’s previous position to the beginning of the line.
User Presses Shift + End → The cursor moves to the end of the line. A selection is formed from the cursor’s previous position to the end of the line.
User Presses Ctrl + C → The selection is copied to the clipboard. The current selection remains.
User Presses Ctrl + X → The selection is copied to the clipboard. It is then deleted, with the cursor then moving to the minimum position of the selection.
User Presses Ctrl + V → If there is a selection, the selection is deleted, and the cursor moves first to the minimum position of the selection. Whatever textual content in the clipboard is inserted at the position of the cursor.
User Presses Textual Character → If there is a selection, the selection is deleted, and the cursor moves first to the minimum position of the selection. A new character is inserted at the position of the cursor. The cursor moves to the position after the newly-inserted character.

That’s a lot! It may seem daunting to support all of these behaviors. But, we can easily support the entire list by finding simple building blocks that are compositional—meaning, they can be combined in many ways—that allow us to express any item in the list.

Let’s break the problem down.

(The remainder of this post is for paid subscribers only)

R. Fleury Computing

UI, Bonus 1: Simple Single-Line Text Input

Comments on this post are for paid subscribers