Random unresponsive cursor areas (case 2931)

[jarvis]

The reason I ask, BTW, is that it appears that the buttons you have created overlap with the primary "corner" regions you have created. Am I getting that right or am I mapping them incorrectly? I'm also going to attach here an excel file that I used to create that map image. That way you can edit it and re-attach it here if you like. Never done this before--seems like a useful way to visualize the regions. (?)

[gerardo]

Wow, thanks for such a quick answer

The regions are those indeed, although region 6 is horizontally centered, like region 5, just wider. If you need me to use the excel file, I'll have to do it tomorrow, as here it's late at night and I should be sleeping

The regions do overlap, but they are never used simultaneusly, for that matter.
The experiment presents a pair of stimuli, randomly distributed among the 4 "corners", and only the 2 respective areas in which images appear are active each time; so the participant can click on one of the 2 images. When selected, the chosen image is shown alone and disappears, and it's then when a "continue" button appears, being the only active area (the only allowed answer). This is region 5. Region 6 is only used once thorough the whole experiment, and when this happens, it offers an alternative option to the one of region 5, with which it appears.

Could the problem be that they overlap when being declared, even if they are never used together in an overlaping way?

The resolution used is 1440x900. On an erlier version, I used 1920x1200 on a different computer. The regions where located proportionally (they where still 340x340, but distributed so they would appear centered on such resolution). I had the same problem as now, but it got solved after I renamed region 5 to 7 and region 3 to 8 (that's it, when I added a new area called 7 with the same attributes as 5, and a new area called 8 copying region 3). I don't know why, but it worked.

It was when I adapted the areas to the 1440x900 screen that the problem started again, this time not being fixed by any renamig trick, as I've tried to describe.

[gerardo]

Thanks to Jarvis comment about the areas overlapping, I got on the right track and fixed the mystery problem; which was, as a matter of fact, that you shouldn't create overlapping areas on the cursor.txt file. I think the manual is a little misleading here, as, at the end of page 69, it states that "you can define as many rectangles of any size and location as you like".

It's also weird, because overlapping area do work sometimes. I started building the experiment with just one area, the continue button. After that, I added the four corner regions, which made the first area stop working, as they overlapped with it. When I created a new area with the same attributes as the first one, but listed after the 4 corner areas, everything went back to working fine. It seems the problem has to do with the size of the area that is blocking the underlaying one; and that Direct gives priority in a reversed order, from last to first, to the areas declared in cursor.txt.

I also had less problems working on a higher resolution (or with smaller areas). With 1920x1200, I could just keep moving the troubled area to the last position, and everything would end up working fine. But when setting 1440x900 on a different computer, the problem kept reocurring.

In short: while overlapping areas seem to work sometimes, the larger the areas or the lower the resolution, the most likely it is you'll end up with unresponsive areas. The way to fix it is to avoid declaring areas that overlap in cursor.txt; which might limit your design, but ensures that everything will work as it should.