Re: [Geotools-gt2-users] Multiscale shapefile

[Jody G. <jga...@re...>]

Emmanuel Pietriga wrote:
> I will, when I'm confident it works. Actually, the current version 
> applies the simplification, but does not save it back on the file 
> system, it uses it directly.
Okay; so that is similar to what shapefilerenderer does right now then.
>> Also your original question was about speed; does the new smaller 
>> shapefile render significantly faster?
>>> It works almost perfectly. I have a problem though. I use the world 
>>> borders shape file referred to in the user guide [1].
>> Right is that this page?
>> - http://docs.codehaus.org/display/GEOTDOC/04+How+to+Read+a+Shapefile
>>
>> As far as I know you are the first person to get  back to me about 
>> that file set. I can try it in udig and see if it has similar 
>> problems...
>>
>> I usually test (and trust) this data set:
>> - http://udig.refractions.net/docs/data-v1_1.zip
>
> Just tried that one's countries.shp. The result was really really 
> weird. Some borders were ok, but many vertices in many more shapes 
> were simply "wild"... see [5].
>
> [5] http://www.lri.fr/~pietriga/2007/12/boundaries4.png
Cool; I saw this last week myself (when trying to switch epsg-hsql) 
notice the dateline giving you trouble !
Try switching to espg-wkt and tell me what you see?

> I'd say there is a good chance I'm doing something wrong, but I don't 
> know what. I simply extract the coordinates from the MultiPolygon 
> (subclass of Geometry), and make them into an actual java.awt.Polygon 
> (with some wrapping around for the zoomable user interface, but 
> nothing fancy). Basically, I draw a straight line between each 
> coordinates, starting from the first point to the last one in the 
> sequence returned by Geometry.getCoordinates()

> Should the list of coordinates actually be handled in a specific 
> manner to draw it correctly as a polygon ?
Somtimes; check if the polygon isValid after your simplification.
Jody