Hi Stefan, I agree this is an important area and deserves being addressed properly. It is also one that would be perfect for a set of unit tests.
I noticed that you started to collect issues that have been filed concerning this topic or related ones, I think this is a very good initiative. A good approach to deal with this topic would be to * create a list of variables that influence the calculation * create a set of combinations of these variables (or eventually a full matrix) with expected results Once this is formulated it should be straightforward to create unit tests, see which calculations are wrong and then fix them while being sure not to break others (because of the test suite). Will you be present in Denmark? Kind regards, Matthias On 05/11/2015 09:57 AM, Blumentrath, Stefan wrote: > > Dear devs, > > With QGIS you delivered a really excellent product with tons of useful > features and I am very fond of it. I am aware that you get lots of > user requests and have to prioritize according to that. Yet, my humble > feeling is that there is one topic which did/does not receive the > attention it probably deserves. > > Proper area and distance calculation is a central functionality of any > GIS. It is used by probably 97.5% of the users and should work in any > case and without the need for the user to define some (at least for > beginners) non-obvious settings, which in addition can get changed > back by QGIS without the user`s notice, e.g. when loading data in > different CRS (including Open Layers). Unfortunately, this seems to be > a long open issue in QGIS (see: https://hub.qgis.org/issues/12057). If > I may make one wish for the Hackfest next week, please make Bug #12057 > (and related issues) a top priority of the event! > > If such basic functionality gives wrong results, personally I would > consider it as a risk for the good reputation of the QGIS project. > Assume a user delivered project results to a customer, which are based > on wrong area calculations (not unlikely that this happens), what do > you think he or she feels about QGIS after that... And even if one > discovers it beforehand. That does not exactly contribute to the > user`s faith in the reliability of QGIS. > > I dare say that this series of bugs is so severe, that if it gets > fixed (hopefully one and for all), it would provide reason enough to > justify a 2.8.3 release (BTW. Thanks for 2.8.2). > > Probably also worth writing a test for this bug, given it`s (long) > history? > > > > Kind regards, > > Stefan > > > > > > _______________________________________________ > Qgis-developer mailing list > Qgis-developer@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/qgis-developer
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer