Steffen Macke wrote:
> > Fmaps can already deal with any multi part object, in the
> > f_<table>_geoobj table just add a tuple with the same tid pointing to
> > the OID of your tabular data <table>
> Does this approach allow a ring polygon?
I'm not sure about it. I have to check within OpenGL, what can be done...
> > About the cache, I'm not sure about it. It may be difficult to program,
> > as each user redraw of the screen will use different selected objects.
> > As for screen redraw when someone move over a window. We could store the
> > display in a bitmap and just redraw this bitmap...
> I think Frank talking about "carried around a tesselated form of the
> polygon in memory" is exactly what I meant with 'cache'.
> OpenEV is there - it's also GPL'ed - I think that Frank will the
> permission to borrow some code (Again, as the shapefile import
> isn't possible without his shapelib).
> Maybe OpenEV has some workarounds for the Mesa tesselation code - or
> it's own tesselation code?
Ok I go for it...
> > About symbology, shapefile does not contain line width, color and
> > others?
> No, all this stuff is stored in the AV project files (Or map document
> files for AI 8).
> Is there support for symbology included in the metadata standards?
There is no support for symbology in the metadata standards, but in
conformance level 2 there is a feature_catalogue_information specification.
I'm coming from the MapInfo World where each object has its own
specification, instead of ArcInfo/ArcView which says that the table/layer
represents the same thing, in the same symbology.