On Fri, Jul 13, 2012 at 08:12:34AM -0700, Paul Norman wrote: > > From: Jochen Topf [mailto:[email protected]] > > Sent: Friday, July 13, 2012 5:02 AM > > To: Paul Norman > > Cc: 'nimix'; [email protected] > > Subject: Re: [OSM-dev] (Multi)Polygon handling > > > > On Fri, Jul 13, 2012 at 03:11:40AM -0700, Paul Norman wrote: > > > For a lot of the invalid cases there won't be a defined output yet. To > > > some extent osm2pgsql is a reference but I have no clue what it would > > > do in some of the possible combinations of invalid tagging and invalid > > > geometry > > > > No problem. Once we do have test cases, we can discuss that here and > > decide that test cases 17 should have result X and test case 23 should > > fail and test case 42 is implementation defined if we can't agree on > > whats supposed to happen. > > The result of anything being left as implementation defined is that > implementations will differ and users will complain when the implementation > they're using differs from others.
Of course we will try to avoid that. After all the whole point of the exercise is to get clarity and comparability between implementations. But if there is disagreement, its the best we can do. And documenting points where we agree to disagree is better than nothing. We'll cross that bridge when we get to it. Jochen -- Jochen Topf [email protected] http://www.remote.org/jochen/ +49-721-388298 _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
