Hi Ragi, I perfectly agree with you. We always pass our data through cleaning procedures, but sometimes it fail. The example I posted wasn't representative, because in that specific case there was effectively a validity problem. But many times the errors happen also with valid data, and that's the case, for example, in unioning. In such cases you cannot know before time what kind of beasts you will get through during the cascaded unioning. So, when I rely on valid datas, it's difficult to explain to the customer why the system fails here and there...
I know I'm facing The Problem, the founmdamental one in geometrical computing with floating point precision, and I know that also commercial softwares suffer of similar problems. I've worked extensively both on ArcGIS and Oracle. I've never said, and I don't think, this is a problem of foss solutions only. giovanni 2010/9/29 Ragi Burhum <[email protected]>: > Giovanni, > I am actually surprised that you bring this comment about a production > environment. When working with real data in production environment, FOSS or > proprietary alike (i.e ESRI), I *always* encounter data that causes Topology > Exceptions (ESRI Geometry/Geoprocessing, GEOS, JTS, whatever) at some point, > it happens. It is just the nature of the beast - data is not perfect. > What I do in such cases, is that before I digest the data, I have a set of > "cleansing" and identification procedures that I put my data through before > putting it in my pretty system. In other words, a mini-fence that I use to > avoid problems (self intersections, nulls, multiparts, whatever). > I apologize if this seems rude, however, the very first thing I do when a > geometry pair gives me problems, is to look at it. That would have shown the > problem the problem right away. > - Ragi >> >> Date: Wed, 29 Sep 2010 17:26:31 +0200 >> From: "G. Allegri" <[email protected]> >> Subject: Re: [geos-devel] TopologyException makes GEOS/JTS very >> difficult to employ in my production environments... >> To: GEOS Development List <[email protected]> >> >> Howard, I know well how foss development works. So, my email wasn't >> meant to ask someone to solve things for me. In this project I'm not >> in the position to make such choices. I've just been able to move >> little money to ask strk a first, basic analysis. I absolutely agree >> with the commercial philosophy behind foss, and when I will be able to >> manage enough money I won't easitate to invest it in that way. >> >> I'm just surprised to see not so many issues raising from this >> topology problems, as GEOS/JTS are maybe the most widespread libraries >> used in the gfoss ecosystem. No problems when using it as a data >> repository, but as soon as I use it massively for geometry processing >> I face, almost everyday, thie TopolyException. So I was simply >> wondering what the others think about this, if they have solved it >> somehow. OS is also sharing best practices, isn'it? I don't want to >> steal industry secrests to anyone, just share ideas... >> >> Giovanni >> > _______________________________________________ geos-devel mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/geos-devel
