Re: [Opengrads-devel] Shapefile extension: first crack

[Brian D. <do...@co...>]

My research into the shapefile format led me to conclude that the  
shaded contour routine in grads will need to be rewritten; Arlindo  
points out some (not all) of the reasons.  I am working on this for  
v2.  We also anticipate implementing shapefile support in grads in  
collaboration with the CPC during 2008... Brian

On Dec 11, 2007, at 4:47 PM, Arlindo da Silva wrote:

> Efren,
>
>   I am back from vacation. I read quickly the ESRI white paper that
> describes the shapefile format. What an awkward format! SVG appears a
> much more elegant approach to vector graphics. In any event, here are
> my thoughts on how to produce shapefiles from within GrADS:
>
> 0) I would produce shapefiles associated with 2D lat-lon "naked
> plots", that is, those without maps, frames or fonts. Anything else
> may not make much sense. (Since GrADS draws its own characters, fonts
> could encoded as polylines but they would be deformed if a map
> transformation is later applied.)
>
> 1) GrADS vector graphics includes both POLYLINES and POLYGON type of
> shapes. Since one can only have one type of shape per shapefile, a
> given image (say contour lines on top of shaded contours) would have 2
> separate shapefiles associated with it: one with POLYGONS, another
> with POLYLINE objects, for a total of 6 files if you factor in the
> .shp, .shx and .dbx subfiles. This is ugly.
>
> 2) There is an annoying requirement on the enconding of polygons: one
> must order the vertices so that the "inside" is to the right of the
> path, something that should have been handled by the renderer.   I
> believe cairo provides functions to determine whether a point is
> inside or outside a closed path, so I may be able to use it.
>
> 3) Each polygon object would have a single part so as to avoid dealing
> with similar vertex ordering issues.
>
> 4) Now, each POLYLINE or POLYGON object will have a color associated
> with it. I am hoping one could save the (r,g,b) values in the
> associated .dbx file, but I am not sure what are the conventions for
> encoding this.
>
> 5) In the case of shaded contours, I am not sure how to encode the
> color bar (mapping of (r,g,b) into field values).
>
>    With all these caveats, would GrADS produced shapefiles still be of
> any interest? I believe Brian has looked into this as well, I'm cc'ing
> him as he may not be subscribed to this list.
>
> Brian: for reference, here is the thread:
>
> http://sourceforge.net/mailarchive/forum.php? 
> thread_name=77fcd6b20711251803j746a965j689063db5d6e4da8% 
> 40mail.gmail.com&forum_name=opengrads-devel
>
>    Arlindo
>
> -- 
> Arlindo da Silva
> da...@al...