Hello,
I have written my own version of a mouse gesture script. I have read thru and briefly tried
deguix's Mouse Gestures and
Jason's version of
nanasisan's (whose identification and naming method is very cool). In fact it was deguix's that proved the reference for mine.
I mostly use my gestures for navigating Windows Explorer in the same way as Opera's gestures (which Wikipedia says it had since April 2001!) and I can't live without them now. My friend says about the same thing
Well, though I think my script has a little better recognition than the other scripts (recognizing a squared 'S' isn't too hard), when trying to distinguish between a 'squared S' and an 'angled S' at the best tolerances, there were difficulties.
Code:
squared and angled S
____ _____
| /
|____ \_____
| \
____| _____/
I think the algorithm and tolerancing have the same inherent weaknesses as the other scripts.
I'd like to hear about other people's experience with mouse gesture scripts:
Are the gestures recognized easily?
With right angles alone you can define around a dozen; Does anyone use more than a dozen commands?
*Would* anyone make use of more than a dozen commands?
With good recognition you might even be able to do simple handwriting recognition (not cursive, and restricting the set to numbers would be better). Could anyone make use of that?
The reason is that there are other ways to recognize a gesture. Better ways it seems. Anyone interested in my writing an attempt?
Thanks for any comments
p.s. Keeping in mind what was said above, aspects of my script:
Gesture definition is similar to deguix's
Gesture algorithm is improved
Gestures are active on an explicitly inclusive basis: specific gestures for specific apps must be included in the definitions and code.
I chose to avoid allowing gestures to function all the time to reduce interference.