Hi Joost, some inline thoughts added.
On 15-05-17 18:49, joost schouppe wrote: > Hi Glenn, > > I'm not sure we should fix this ourselves. I think we can ask Be-mobile > to fix this. But maybe a revert is in order. That would also be a lot of work btw, very small changesets. I don't really like reverts (unchecked) as it tends to throw away the baby with the water. > Be-mobile didn't discuss this in advance, but just let us know they did > afterwards. I personally failed in following this up, so now we have > this mess on the mailing list. First steps are always hard, and when you > try to do too much in your first step, you can have a big mess up. Let's > not judge them too much, and try to make this a learning experience, not > a first and last step. But there is definitely a lot of learning to be had! I agree, I tend not to fix the blame, I rather fix the problem. > I can't judge the quality of their WMS, as I can only get it to tell me > "something is wrong here", not what needs to happen. The base data from > AWV seems to be correct, at least in the case Marc found. So I would > guess it's a question of instructions gone wrong. If I had to guess, that would also be mine, perhaps communication problem to the person doing the work. > In many cases there is both AWV data (which the people at AWV were quite > confident about) and Mapillary traffic signs, so we do have some > material to work with. It would be nice to have an idea of the size of > the problem, so we can give more info to Be-Mobile. Of course we need a > quick solution, but I think it would be best if a larger problem caused > by a paid mapper doesn't need to be fixed by a volunteer. I'll try and make a Q/A tool for this, if I have the datasets. I'm messing about with the GRB dataset again, might throw in the roads as well. primary/secondary roads are a relatively small dataset. As long as the problems isn't the source data, we can detect those. Wouldn't it be nice if we got paid for edit work ? :) Glenn _______________________________________________ Talk-be mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-be
