I was wondering, what do you think (interpret this only as a question) about introducing validation in iD in the future?
Using MAGIC integrated circuit design tool, that does DRC (Design Rule Check) in real time and highlights errors inspired me that OSM editors could also incorporate this. It makes sense to me. First, users will not be overwhelmed by a sh**load of errors at once and second, they could learn what they actually do wrong. But this poses challenges, because sometimes when you're editing, there will be a temporary error state, to disappear just when you finish a sequence (e.g. you don't enter all tags at once so there will be a transient "place of worship without religion" error.) That type of error message should not happen, because spamming irrelevant errors only makes users ignore them. Still, there are checks that can be safely made in real time, like all sorts of geometrical tests (self-intersections, building crossing another building and so on.). Maybe good-enough heuristics could be applied for when the user stopped editing a feature and moved on to another, to address the "temporary error" issue. Anyway, thanks in advance to anyone who makes iD more iDiot-proof. It really matters a lot, for example there was press coverage (Polish News Agency) back in 2014-08-18 that generated 500 or more new users who obviously contributed a fair share of mistakes. There was simply no manpower available to check edits of all the users, let alone message them on what they did wrong. Having no severe errors is quite a point of honor to me, as I think we must try to be free of all these "that cursed satnav told me to do this" situations. Steve Jobs once told something along the lines of "We don't ship junk. We make products that we could recommend to our family and friends" and surely anyone tech-savvy can relate to that feeling of embarrassment when a cool gadget/software you show to your family happens to betray you. Have your navigation lead you off-road (see: wrong road tagging), people will tell that "OSM is shit" even though other map products are not ideal as well. Michał
_______________________________________________ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
