Here is an interesting mystery sure to drive someone nuts eventually. It seems like it is some kind of strange interaction between d.grid and our rendering algorithms–which have been tested against raster map cells.
impact_wxp1.png shows how a grid overlays single raster cells at high resolution. No matter how I shift the map around with the pan tool or zoom tools, the relationship between the cell and grid stays the same–aligned in the E-W direction and misaligned in the N-S direction. Furthermore, if I use the cursor to view the coordinates in the bottom status bar, the E-W grid lines and cell borders match what is showing for easting coordinates in the status bar when I align the cursor to the E-W grid/cell border. However, in the N-S dimension, the location of the cursor does not match the grid line position, but is about 25% of a cell above it. I don’t know if it matches a raster cell border because I don’t know what the border value is supposed to be.
There is also an error printed (probably unrelated): " Type of layer is not supported yet" every time a map is refreshed showing a grid, although the grid displays without problems.
Turning to the TclTk display, for all zooming in and out, cells, grid, and cursor position all match up perfectly (impact_tcltk1.png). However, if I subsequently pan around, the cell display and grid no longer align, though in this case the cursor position exactly matches the gridline values in both E-W and N-S dimensions.
Michael
Michael Barton, Professor of Anthropology
School of Human Evolution & Social Change
Center for Social Dynamics & Complexity
Arizona State University
phone: 480-965-6213
fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton
(attachments)
