Detect It
Detect if a device is mouseOnly
, touchOnly
, or hybrid
, and if the primary input is mouse
or touch
. Also detects if the browser supports the Pointer Events API, the Touch Events API, and passive event listeners. Detect It is tree-shakable and side-effect free.
This is useful for creating device responsive UX and responding to user interactions. When creating apps with device responsive UX it is important to know what the user can do. Can they hover? Can they swipe? Etc. Once it's known what the user can do, the next question is how does the app listen for user input. For example, it's not enough to know that the user is on a touch device, the app also needs to know how to listen for touch input, should the app set PointerEvent listeners or TouchEvent listeners? For more on this see the Recommended usage section.
Live detection test (code in the demo repo)
detect-it
v4 is currently in pre-release, use the @next
tag to install it, v3 is available here
npm install --save detect-it@next
import * as detectIt from 'detect-it';
import {
deviceType,
primaryInput,
supportsPointerEvents,
supportsTouchEvents,
supportsPassiveEvents,
} from 'detect-it';
deviceType: 'mouseOnly' | 'touchOnly' | 'hybrid';
primaryInput: 'mouse' | 'touch';
supportsPointerEvents: boolean;
supportsTouchEvents: boolean;
supportsPassiveEvents: boolean;
deviceType
mouseOnly
| touchOnly
| hybrid
Indicates if the the device is mouseOnly
, touchOnly
or hybrid
. For info on how the detection works and how specific devices are classified see Notes on detecting deviceType
.
import { deviceType } from 'detect-it';
if (deviceType === 'hybrid') {
}
primaryInput
mouse
| touch
Indicates if the primary input for the device is mouse
or touch
. For more info on how to use primaryInput
see the Recommended usage section.
import { primaryInput } from 'detect-it';
if (primaryInput === 'touch') {
} else {
}
supportsPointerEvents
boolean
Indicates if the browser supports the Pointer Events API. See MDN's Pointer Events and the W3C Pointer Events specification for more information on Pointer Events. See Can I use for current support.
import { supportsPointerEvents } from 'detect-it';
if (supportsPointerEvents) {
element.addEventListener('pointerenter', handlePointerEnter, false);
}
supportsTouchEvents
boolean
Indicates if the browser supports the Touch Events API. See MDN's Touch Events and the W3C Touch Events specification for more information on Touch Events.
import { supportsTouchEvents } from 'detect-it';
if (supportsTouchEvents) {
element.addEventListener('touchstart', handleTouchStart, false);
}
supportsPassiveEvents
boolean
Indicates if the browser supports passive event listeners. See this Passive Events Explainer for more information on passive events. See Can I use for current support.
import { supportsPassiveEvents } from 'detect-it';
if (supportsPassiveEvents) {
document.addEventListener('scroll', handleScroll, { capture: false, passive: true });
} else {
document.addEventListener('scroll', handleScroll, false);
}
Pre-built option served from Unpkg CDN
Optionally, instead of using npm install
you can load Detect It directly in the browser. A minified UMD version is available from Unpkg for this purpose.
<script src="https://unpkg.com/detect-it/dist/detect-it.umd.min.js"></script>
if (window.detectIt.primaryInput === 'touch') {
}
Recommended usage
TL;DR:
- Use
primaryInput
for creating device responsive UX that optimizes the user experience for either mouse
or touch
input (note that the app should still be usable by both inputs). Use this along with classic responsive design that adapts to screen/window size to create a fully device responsive app. - Listening for user interactions:
- If the browser
supportsPointerEvents
then only set PointerEvent listeners and use pointerType
to determine if the interaction was from mouse
or touch
. - Otherwise always set both MouseEvent and TouchEvent listeners and use
event-from
to ignore MouseEvents generated from touch input.
Device responsive UX
Device responsive UX is about creating web apps that feel native on every device. This goes beyond classic responsive design, which only responds to the screen/window size, and includes how the user can interact with the app (the capabilities of the device). Can the user hover, swipe, long press, etc?
There are 3 parts of device responsive UX: Size (size of screen/window), Can (what the user can do/capabilities of the device), and Is (is the user hovering, touching, etc). Size and Can need to be known at render time (when the UI is rendered), and Is needs to be known at interaction time (when the user is interacting with the app).
- Size
- This can be determined using media queries, for example
(max-width: 600px)
, either applied via CSS or in JavaScript by using something like react-media
.
- Can
- This is what Detect It is built for - knowing at render time what the capabilities of the device are (what can the user do). There are a number of ways that you could use
deviceType
or primaryInput
to optimize the UX for the capabilities of the device, however, in most cases I've found it makes sense to just use primaryInput
and optimize the UX for mouse
or touch
, while ensuring that the app is still usable by both inputs.
- Putting Size and Can together, my preferred approach is to optimize the UX for 4 scenarios:
- Wide screen with
primaryInput
mouse
: desktop/laptop with a normal window - Narrow screen and
primaryInput
mouse
: desktop/laptop with a narrow window - Wide screen with
primaryInput
touch
: tablet - Narrow screen with
primaryInput
touch
: phone
- Is
- Is the user hovering, touching, etc. To help with this I created React Interactive which provides a callback for interactive state changes (
hover
, mouseActive
, touchActive
, keyActive
) and allows you to style touch interactions in ways that are not possible with CSS pseudo selectors.
Setting event listeners
Setting event listeners can be thought of as either setting PointerEvent listeners or setting MouseEvent and TouchEvent listeners. PointerEvents can do everything that MouseEvents and TouchEvents can do (and more), without having to worry about if a MouseEvent was caused by touch input and so should be ignored. It is recommended to use PointerEvents if they are supported.
PointerEvent listeners
If the browser supportsPointerEvents
then only set PointerEvent listeners and use pointerType
to determine if the interaction was from mouse
or touch
.
import { supportsPointerEvents } from 'detect-it';
const handlePointerEnter = (e) => {
if (e.pointerType === 'mouse') {
} else {
}
};
if (supportsPointerEvents) {
element.addEventListener('pointerenter', handlePointerEnter, false);
} else {
}
MouseEvent and TouchEvent listeners
If the browser doesn't support PointerEvents, then there are a couple of ways to approach setting mouse and touch event listeners.
Note that a touch interaction will fire TouchEvents as the interaction is in progress (finger on the screen), and then will fire MouseEvents after the touch interaction has finished (after the finger is removed from the screen) to support sites that only listen for MouseEvents.
Option 1: If the device is mouseOnly
or touchOnly
then only set mouse or touch listeners, and if the device is hybrid
set both mouse and touch event listeners and ignore MouseEvents caused by touch input (you can use event-from
for this).
Option 2: Always set both mouse and touch event listeners and use event-from
to ignore MouseEvents from touch input.
I prefer option 2 as it's simpler to code and I haven't noticed any performance impact from setting extra listeners (note that setting TouchEvent listeners on a browser that doesn't support TouchEvents is fine, the browser will just ignore the event listeners).
import { supportsPointerEvents } from 'detect-it';
import { eventFrom } from 'event-from';
const handleMouseEnter = (e) => {
if (eventFrom(e) !== 'mouse') return;
};
const handleTouchStart = (e) => {
};
if (supportsPointerEvents) {
} else {
element.addEventListener('mouseenter', handleMouseEnter, false);
element.addEventListener('touchstart', handleTouchStart, false);
}
Notes on detecting deviceType
To determine the deviceType
and primaryInput
Detect It uses multiple API detections and media query results to triangulate what type of device is being used. The entire detection is done when the script is imported so the results are known at render time (Detect It doesn't set any event listeners).
Detect It has a wide definition for what constitutes a hybrid
device, or rather a strict definition for what are mouseOnly
and touchOnly
devices, because if a device strays from only a fine point and hover with a mouse, or a coarse touch with a finger, then it should be treated uniquely when considering how the user will interact with it. Below is the source code for determining deviceType
:
export const deviceType =
hasTouch && (hasAnyHoverOrAnyFinePointer || !hasCoarsePrimaryPointer)
? 'hybrid'
: hasTouch
? 'touchOnly'
: 'mouseOnly';
Some hybrid
device examples
- A touch capable Chromebook
- A touch capable Windows computer (both when it's used as a regular computer, and when in tablet mode, e.g. Microsoft Surface without a keyboard)
- A Samsung Galaxy Note with stylus
- All iPads now that they support a mouse and keyboard (note that Apple makes it impossible to know if a mouse or keyboard is attached, so iPads are always treated as a
hybrid
with primaryInput
touch
)
Detection limitations
These detections are limited by how the browser presents itself (the APIs it exposes and how it responds to media queries) so there is the potential for errors. As anyone who has done web development knows, browsers are quirky and they don't always behave in expected ways, so Detect It uses real world browser behavior to triangulate the deviceType
and primaryInput
. For more on this see the comments in the source code for notes about detecting the device type. Also, tests have been written to mock a number of different browser behaviors and edge cases. Cloning this repo and running npm test
will provide insight into how detect-it
will preform on different devices.