I think there are two big issues with the correction procedure as currently implemented. Firstly, the mouse-based manual correction is very fiddly and I think there's a much better way to implement this. When I previously did manual corrections, I basically just looked at each fixation one-by-one and typed the line number. So, I would just end up typing, for example, 111111122222222222233333333... etc. This makes the process really quick easy because you don't need to make fine mouse movements (which is uncomfortable for long periods). Secondly, and perhaps more importantly, I think it's really important that the user can see the upcoming fixations while they are working, but Fix8 hides this information from the user during the correction process. This is crucial because it's only by seeing the entire fixation sequence as a whole that one can appreciate how the fixations are drifting.
So, basically, the way I think it should work in Fix8 is that all fixations are shown the whole time, and the software then highlights them, one-by-one, and the user just types the line number (or the up/down arrow keys) to snap each fixation to a particular line (of course, you could also allow mouse functionality too). The previous fixations would be in one color (e.g. red). The upcoming fixations would be in another color (e.g. blue). And the current fixation would be a third color (e.g. black). You could also show the line numbers at the side of the plot, and if there are more than 9 lines, you could use the letters A, B, C... etc. The same should be true of the assisted correction: it's also important that the user can see the upcoming fixations to help them decide if the algorithmic recommendation is correct. I know that it's currently possible to move the progress bar to look ahead, but this is quite cumbersome and error-prone because you have to keep looking ahead and then moving the progress bar back to find where you were. Or you have to look at all the fixations and try to remember what the drift looks like while performing the correction.
I think there are two big issues with the correction procedure as currently implemented. Firstly, the mouse-based manual correction is very fiddly and I think there's a much better way to implement this. When I previously did manual corrections, I basically just looked at each fixation one-by-one and typed the line number. So, I would just end up typing, for example, 111111122222222222233333333... etc. This makes the process really quick easy because you don't need to make fine mouse movements (which is uncomfortable for long periods). Secondly, and perhaps more importantly, I think it's really important that the user can see the upcoming fixations while they are working, but Fix8 hides this information from the user during the correction process. This is crucial because it's only by seeing the entire fixation sequence as a whole that one can appreciate how the fixations are drifting.
So, basically, the way I think it should work in Fix8 is that all fixations are shown the whole time, and the software then highlights them, one-by-one, and the user just types the line number (or the up/down arrow keys) to snap each fixation to a particular line (of course, you could also allow mouse functionality too). The previous fixations would be in one color (e.g. red). The upcoming fixations would be in another color (e.g. blue). And the current fixation would be a third color (e.g. black). You could also show the line numbers at the side of the plot, and if there are more than 9 lines, you could use the letters A, B, C... etc. The same should be true of the assisted correction: it's also important that the user can see the upcoming fixations to help them decide if the algorithmic recommendation is correct. I know that it's currently possible to move the progress bar to look ahead, but this is quite cumbersome and error-prone because you have to keep looking ahead and then moving the progress bar back to find where you were. Or you have to look at all the fixations and try to remember what the drift looks like while performing the correction.