help!--v.digit to raster conversion

[Lars Schylberg]

>   Perhaps think an image dithering technique might be more general, ie
>   d.rast -d roads (-d means use dithering).
>   One way to find out is to run p.map to the preview driver and see if the
>   roads show up better.

Maybe I don't know enough about dithering techniques, but as far as I
understand about image dithering techniques, the idea is to use more
pixels to describe one cell in a better way.  But the idea I had is to
describe many cells in a raster data base by selecting the most proper
representation of them among the cells that are under the super cell that
should be presented as a pixel on the screen.

d.rast -mp roads water open_fields ( -m mode selection -p priority roads 
                                       water open_fields any other cats )

I am not sure about how it works today in d.rast. But as far as I
understand is, that if I have raster data in 5 m resolution and have a
region and frame setting on my screen that allows each pixel to be
repressented by a 50 m pixel, then it selects the first 5 m cell in
the upper row and column of the 50 m supercell. then it skips 50 m
forward and picks next cell value and so on.  It does the same row
wise, that it skips 50 m down the rows and picks the next row there.

Please correct me if I'm wrong !.

I tried to look at this once in the source code but I'm not sure if
understood everything right, since there so many pointer operations
where involved in this part where this action takes place.  I think
this would be of general interest to have this clearified in detail
since this is why linear objects can behave strange sometimes when
they are displayed.

Lars

Lars Schylberg                 Email: larss at fmi.kth.se    (Internet)
Department of Photogrammetry          larss at sekth.bitnet  (Bitnet)
Royal Institute of Technology  Tel.   +46 8 790 86 33     (office)
S-100 44  STOCKHOLM, SWEDEN    Fax.   +46 8 790 66 10