I've got a propositon for a pinch-gesture for single-touch screen.

We can call it "tap and swipe"

You tap once on the screen (fixing the center of the movement).

Then (within half a second) you make a swiping movement from a starting point to an end point.


The transformation is defined by two vectors:
- the vector of the center and the beginning point
- the vector of the center and the end point

This is a combination of a rotational matrix and a stretching matrix that is so easy to calculate that you can write it down immediately.



Even better:
Matrices about the same fixed point are easily added linearily.

Therefore: If the swiping movement is not only a straight line but a curve also the components of the matrices describe the same curve. You can extrapolate the mouvement. The impetus of the mouvement can be continued a little bit during the time.


Even better:
Even a 3D mouvement is possible with a pressure sensitive touch-screen: bringing elements more in front or back.


The "Tap and Swipe" gesture can be continued by a "Swipe and swipe" gesture taking the end point of the first swiping gesture as center of mouvement for the second swiping gesture.

A great idea! I wonder if the tap could be misinterpreted, and whether it should be a double-tap instead. But on reflection, I think a single-tap is okay.

Tap-swipe. I like that.

Seems like that would be difficult to distinguish from just a swipe. I find the rocker switch great for zooming. I prefer to keep my fingers from the screen; the stylus and switches work great and its hard to see around my fingers.

Me too!

I don't like to put my fingers on the screen.

The swipe gesture can also be realized by a stylus.

Already now the device is capable to distinguish a swipe gesture from a tap gesture.
And with "tap and swipe" the tap and the swipe gesture are even locally separated.

There isn't even much programming to be done. The QT-Framework already provides a "tap" and "swipe" gesture and presents structures to register custom gestures. The only thing we have to do is to emit an artificial "pinch gesture"message when a "tap and swipe" gesture is recognized.


In this way we emulate the "pinch gesture" by a "tap and swipe" gesture.


But this has to be done in the GUI (because I don't want to program it for each application).

There should be also an entry in the preferences to allow the activation of the emulation of "pinch gestures" by "tap and swipe gestures".



The problem is what "kwtm" speaks about.

If the application provides already a message handling for the tap gesture the device can decide only after half a second if we are treating about a pure tap gesture or about a "tap and swipe".
But sometime bothes gestures can exist together. In this case no waiting is needed.

The cw/ccw-screwing-gesture works pretty good, I think, already.
The problem is, it works only in microb but not in other applications (like in xterm, drawing, Fennec, ...)

When meego will have multitouch-gestures, the pinch-zoom-gesture will be the most used one. For running Meego-apps in N900 some gesture has to be mapped to pinch-gesture. The "tap and swipe" could be a good alternative to the cw/ccw-gesture.

(Or actually, I believe there is just "zoom-gesture" event received to the apps and the lower level GUI-system maps different kind of gestures to this "zoom-gesture"?)

With 3D-objects, "tap and swipe" could be much likely be rather 3D-rotation than zoom, because it can have the direction. cw/ccw-gesture works intuitively only on z-axis. With 3D, there has to be a way to distinguish translation-gesture from rotation-gesture. Simple swipe for translation, and tap-swipe for rotation would sound suitable.