It seems to work with this little Basic program:
10 PRINT JOY(1)
20 GOTO 10
Then I have tested the mouse with the EGI, to discover that one of the pulleys is a lightly eccentric, so the vertical coordinate doesn't work very well.
Also the resolution is very low, I have to lift the mouse two times to make one fast diagonal. But I have also noticed that I don't have to lift the mouse with a slower movement. I have to move at its rhythm...
It can be that on a text environment it is enough.
I want to know if a 32mm ball could fill better the hole. Tomorrow I will buy some more gum washers to find the perfect pulley to continue the tests.