Dear list, I am a huge fan of the new metafun interface. Hans asked for luametafun requests, so here are a few. - color by value in lmt_surface We want to show a surface plot of a function where the color indicates the function value (instead, or in addition to a light source). The easiest would be to have variable cf for the brightness factor (as it is), and in addition variables such as zt, minz, maxz, akin to the color formatter for lmt_contour. This becomes handy when colors are to be aligned with function values (e.g. negative = blue, zero = white, positive = red, see MWE attached). - controlling line levels in lmt_contour It would be convenient if certain levels could be set by the user. E.g. to make sure the zero level contour line is included. The user interface could potentially take a list, levels = { -1, -.5, 0, .5, 1 }, of contour level lines. See MWE attached. From my understanding marching squares can do that. A possible reference implementation could be QuadContourGenerator [1] in matplotlib, a bit of description can be found in [2]. - colormaps Contour plots often use rather complex colormaps, such as jet, rainbow or veridis. They are optimized for perceptual resolution, color blindness, screen/print or certain use cases. The functionality is already there, but from a user's perspective, it would be much easier if some of these were provided by the interface. The interface could be color = "colormap('jet')(l/n)". Indeed, these are useful in other use cases as well, such as bar charts, surface plots or shades. One rudimentary example is in the attached file. - number of levels in lmt_contour (potential bug) The number of levels is different from the user's input. Depending on the output range it can be lower or higher (see MWE attached). Cheers, Jacob Kauffmann [1] https://github.com/matplotlib/matplotlib/blob/96ff1146ad053c5c7481efd9de039c... [2] https://github.com/matplotlib/matplotlib/blob/96ff1146ad053c5c7481efd9de039c...