The only thing that the browser gives to your Javascript application
is a sequence of touch events (touchstart
, touchmove
, touchend
,
touchcancel
). Basically, when your finger touches the screen you get
touchstart
, when the finger moves you get touchmove
, and when you
lift the finger from the screen you get touchend
.
So, when you do any gesture such as tap, swipe or pinch, your
application will only get a sequence of touch events. E.g., tap could
be just touchstart
+ touchend
that happen within a short period of
time. Swipe is touchstart
+ many touchmove
’s with increasing
velocity + touchend
, and so on.
Browser does not give you any events that directly correspond to
gestures, something like tap
, swipe
or pinch
.
Your Javascript application must recognize gestures from the sequence of touch events. Basically, you need to implement a function that analyzes the last event that happened and the sequence of preceding events (with timing information) and decides if this looks like a certain gesture.
For example, if it is touchstart
+ touchend
and less than 100ms
passed between two events then it looks like a tap. You can imagine
what it takes to detect more complicated gestures such as swipe (that
can happen in a different directions).
HammerJS is just one of the libraries that does exactly that: recognizing gestures from the sequence of touch events.
However, browsers actually do recognize gestures, they just do not share this information with you. This is disappointing, if you think about that: browsers have a lot of code that understands how the device works and how operating system detects gestures. But because there is no accepted standard for that, you have to detect gestures manually, as described above.
When the browser detects gesture in the sequence of touch events, it normally does something that makes sense in this situation.
For example, tap generates the same event as mouse click would:
click
. Pinch-zoom allows you zoom the page for easier reading.
Vertical panning leads to scrolling the page.
If the web page does not know anything about touch events (or does not bother) then it’s going to work on touch devices in a way that people expect.
Your application can actually disable any default browser behavior if it thinks that this would be better. Even the scrolling can be disabled, even tapping on links.
Technically this is done either by calling event.preventDefault()
on
one of the touch events, or by specifying certain CSS properties (see
below).
As the simplest example, if you listen for touchstart
event and call
preventDefault()
on all of them then your page will become
completely unresponsive on touch-based devices.
You have to understand very well what exactly gets disabled. Suppose
that you have a laptop with touchscreen and a mouse. You can take any
<a>
tag and listen on touchstart
event on this DOM element,
calling preventDefault()
on it. This will prevent default browser
behavior and stop generating click
events.
If you tap on this element using touch screen, nothing will happen.
However, if you use the mouse and click on the link, it will work,
because mouse click generates click
event directly.