Anke, I think you might be trying to push your computer beyond the limits of reasonable floating point number arithmetic. I don't have the expertise in this subject that other programmers on this list do, but I would recommend the following:
- Read a little bit about floating point numbers and how they work. I have actually found the wikipedia page has the info that is easiest to understand. - Set up your code to test for results within a certain tolerance. For example: Don't compare exact coordinate values, but see if coordinate values are within a range. Don't ask if two coordinates are exactly the same, but rather if they fall within a bubble of some given radius. The same thing can be done if you want to know if a coordinate falls on a line. Don't ask if the point is exactly on the line, but if it falls within a given buffer of the line. At some point, the errors get so small they don't matter. If they do matter, then you need to use something besides floating point numbers, and that makes your math a lot slower. :] If you are morbidly interested, I started work on some code that allowed the client code to set the level of precision for arithmetic calculations. A person could set up a geometry library using this code with their required level of precision. I never got that far with my experiment, but that was my eventual goal. I was always curious just how much slower it would be. :] The Sunburned Surveyor On Tue, May 26, 2009 at 5:28 AM, Anke Trittenbach <[email protected]> wrote: > Thank you Matthias for the information, these have helped me already a lot. > > Now I have tried about a PrecisionModell that the Linestrings intersect the > generated Point of the Intersection. The PrecisionModell was set that the > geometries had no decimal places. However, this time the Linestrings do not > intersect the point from the intersection. I thought the PrecisionModel is > for such problems?! > > -- > Neu: GMX FreeDSL Komplettanschluss mit DSL 6.000 Flatrate + Telefonanschluss > für nur 17,95 Euro/mtl.!* http://portal.gmx.net/de/go/dsl02 > _______________________________________________ > jts-devel mailing list > [email protected] > http://lists.refractions.net/mailman/listinfo/jts-devel > _______________________________________________ jts-devel mailing list [email protected] http://lists.refractions.net/mailman/listinfo/jts-devel
