fokijuicy.blogg.se

1. #Engauge digitizer help with defining axis software#
2. #Engauge digitizer help with defining axis code#

Do you think the original request also has merit (to echo the entered coordinates somewhere)? For a while, it seemed like it wasn't accepting scientific notation and I had no way to confirm this. To aid reliability, it is recommended that two researchers carry out data. It wasn't really clear what I was supposed to take from that.

#Engauge digitizer help with defining axis software#

So, the software should be able to take the plot coordinates, transform to log coordinates, do everything else it needs, and then untransform back to graph coordinates when I request output. I've nominally already indicated that my axes are log-scaled. Taking this approach, I'd rather seen Engauge do the rescaling internally. Would that work for you? Maybe using log scale? An immediate fix is to manually rescale your data - like maybe adding 20 to each exponent in the y coordinate. As a point of reference, it took months to get log scale working well. Machine precision is ~1e-15, which wouldn't cover the ~1e-20 calculated for the determinant.Īlthough all the issues could be fixed, this would result in a huge amount of work. Besides, would newer versions of Qt even have a different, useful, criterion? The one cited is 1e-12 (if I counted correctly).

#Engauge digitizer help with defining axis code#

Although this is where the problem currently manifests, I am sure that somehow getting past that point by upgrading the code would immediately cause other problems elsewhere in the code. To determine if the matrix is invertible, the determinant's value of 9.9e-21 is compared in qglobal.h to 0.000000000001. So, the software should be able to take the plot coordinates, transform to log coordinates, do. Specifically, the 3x3 matrix used to convert between screen and graph coordinates uses (x0,y0,1) as the first column, (x1,y1,1) as the second column and (x2,y2,1) as the third column. Ive nominally already indicated that my axes are log-scaled. The Qt library, on which Engauge runs, is not happy with the 27 orders of magnitude in the coordinates. In any case, before closing this issue I would like to, at least, improve the error message to explain when the issue is a huge dynamic range in the coordinate values. As a point of reference, it took months to get log scale working well.Īn immediate fix is to manually rescale your data - like maybe adding 20 to each exponent in the y coordinate. Although all the issues could be fixed, this would result in a huge amount of work. Specifically, the 3x3 matrix used to convert between screen and graph coordinates uses (x0,y0,1) as the first column, (x1,y1,1) as the second column and (x2,y2,1) as the third column.