I am implementing a fix for the incorrect MinY of the Geometric Image Analysis in Python, as discussed in this post. While doing so, I noticed an odd behavior. It is allowed to specify the Image Size as negative for the Geometric Image Analysis, but this causes problems with the resulting output. With the example Double Gauss 28 degrees field, I can create the following analyses using the API with slightly different settings:

Showing that the extent of the axis with a negative image size and the number of pixels set to 100 will always equal 100. First I thought this was an API problem, but this is the same output in OpticStudio:

Note that the image width of Analysis 2 is 100 millimeters while it should be (-)50 millimeters:

Increasing the number of pixels to 200 increases the field width of the analyses with negative image sizes to 200 millimeters while it should be (-)50 millimeter and (-) 100 millimeter.

Is the behavior where a negative image size results in a field width equal to the number of pixels by design? Or should we not be able to specify it as a negative number in the first place?