As I understand, Cerupcat is working a new mouse driver… but I just wanted to bounce this idea off everyone to see if its even a decent idea.

I know the basics of c++ and c#, and have studied the touchlib a bit… and from what I understand touchlib detects and sends the blob data to other programs.

There has been a lot of talk about gestures, standardizing them ect. but a main problem I see with all of our setups so far is that the multitouch is only really useful in our custom programs…

My solution… Still in the idea stage.

Create a program X that runs in the background and recieves the messages from touchlib. Just as the mouse driver does.

This program X would then load a config XML file… this XML file would tell the progam X which custom gestures (or standard gestures) do what in each program.

Program X would then perform a loop

Get the Current Window.
Determine which gesture was made.
See if the gesture and the program match up.
if they do…
Send a custom window message to the program.

Some C# ideas on how to perform these tasks

To get the current window:

To send the Messages to the window… although can’t C# messages be sent as an event?

The advantage to doing this would be to use multitouch gestures in everyday programs like Mozilla… its almost as thes gestures act as hotkeys.

Feedback on this idea is appreciated..

I like that idea!

If there were a gesture in the mousedriver for an alternate mouseclick, that alone would cover a lot of applications and make them MT friendly.
It would be a bonus to be able to work with programs that are mostly mouse clicks and drags. Photoshop for instance.
Another cool integration into this program X would be a gesture to bring up a keyboard. Text and numerical input would be great.

If I had a mousedriver like that, I would probably use it and my MT table for a lot of my Photoshop and Illustrator work.

Similar mouse driver (no gestures) already exists in MultiTouchVista framework. It can be extended to support gestures.

Here’s a link to an interesting Toolkit done by the University of Calgary.  It’s a toolkit that allows you to add multiple mouse cursors into the current OS environment.  Dunno if its of any use, but I guess its a good starting point.

http://grouplab.cpsc.ucalgary.ca/software/SDGT/index.html

Progen - 12 July 2008 06:49 AM

Here’s a link to an interesting Toolkit done by the University of Calgary.  It’s a toolkit that allows you to add multiple mouse cursors into the current OS environment.  Dunno if its of any use, but I guess its a good starting point.

http://grouplab.cpsc.ucalgary.ca/software/SDGT/index.html

It uses same technology (Raw Input API) as MultipleMice input provider in MultiTouchVista.

I’m writing(READ: Hacking together in as little time as possible) a thing like that in Python.  I will let you all know when I get it working.

hmm, even better would be a custom gesture creater, that allows the user to create gestures and assign them to mouse or keyboard clicks, Much like most opperating systems it is possible to configure the mouse in a miriad of different ways. This allows the user freedom rather then conforming to a standard that hasnt even been set in stone yet. Also no repeated stroke is exactly the same to the program needs a threshold, Perhaps repeating the gesture 3 times and the program takes the average. The user interface could replay your gesture as a blob or something more fancy. Until your happy obviously a few standard gesture forms would help get started but tweekability will suit the more picky user!

Perhaps an international gesture standards agency (IGSA!) needs to be set up so everyone can sweep, swipe, jib, jab, pinch, punch, roll etc in a simalar way!

anyway im sure this has probably been discussed in another post but just a thought!