I hope that everyone already knows that I am building a character recognition magic tool as part of GSOC 2008. I have created the part where characters drawn by using the pen tool are converted to .ink files and are recognized as language alphabets. This is done through shape recognizer and hence is completely language (not programming language, but spoken language) independent.

But now I am building the user interface for my magic tool and I am encountering few usability issues. I would like some hints and suggestions from other tuxpaint developers. After the user draws a character using the pen tool, on what event should the recognition begin? Bill had mentioned that the user should click on checkbox/button when done. But I have not found an easy way to create a checkbox/button using magic tools. Another Idea could be that when another magic tool is selected the recognition should begin i.e. the user first selects my magic tool called drawText, then draws a character on the canvas, then clicks on another magic tool, then the recognition will begin and the character will be displayed as font. 

But even a magic_tool_change_event is unavailable through the api. How do you suggest I can implement this workflow? Any other ideas?  


Namrata Nehete.