What’s the issue when OSMOSE for example shows?: river deviation of 169.0m
It refers to the angle of the bend. The number is an interpretation that the way it was mapped is possibly incorrect. I’ve corrected many streams and rivers which were mapped very crude, so retraced long sections with more refinement (with newer imagery).
Ok, thanks for the answer.
I have a bunch of these types of errors, but think most of those are false positives, but I shall look.
Osmosis (as well as other programs) alerts the last person who edited an object. It does not mean that the person is responsible for the problem.
At the beginning (months ago) I thought it was strange, but today I don’t care anymore, because in the case of humans who edit, the tendency is for people to correct other “osmosis alerts” in the same region. So the end result is positive! Less people auto-editing, more care!
About this specific alert for rivers: it’s not a false positive. For the most part, the river was actually drawn quite crudely.
So yes, if you touch a river, your user will be the last editor of the river, so Osmosis will alert you to that.
As an non-English speaker I have a very vague understanding of how to interpret “quite crudely”.
The issues osmose flagged were about a major river in Sweden (Ångermanälven – Wikipedia), which is long (463 km). Maybe the river should have been mapped with more nodes, making bendings more smooth. I don’t know, I was not the person how added the river. But I do not see any wrong with the current mapping, and I regard these osmose issues as false positives.
Wow. Now I get your point!
The rivers triggering the Osmose warning in Sweden (compared to what I’m most accustomed to) are very wide ones, so my comment (likely even the default osmose help message) will make less sense. So I think it is better to clarify the “quite crudely" because even here, wide rivers such as Uruguay River, the recommendation to review is not intuitive to be done only by visuals: these kinds of river as so huge, that user could zoom out. However, the typical river (at least in the regions I map that trigger this warning are more like 2 to 4 meters wide.
The good thing is the following: I assume that despite the Ångermanälven being 400 km long, the bare minimum to just fix the points of this warning would not take proportionally a lot of time. I think the close I done was the Río Megaruma in Mozambique, initially was a 200 km OSM way, them I splitted in two (changesets 134837939 and 134838516) and I think I took around 2 hours, maybe less if my browser did not crashed when iD allowed me to try submit over 2.000 nodes in a single way, so server refused. It’s a bit boring to fix such rivers, but if done in one go, the complaints stop for everyone. So I suspect if you learn how to do this, you are likely to do it for others of the entire country.
So, if no one else explains a bit better how to pass this kind of Osmose check on this forum thread, maybe in the next few days I will give some print screens with annotations how I do it. But the strategy for wider rivers and the smaller ones likely to be different, because on the smaller ones often Osmose warnings are a proxy to issues on the entire part of the river, while on the big ones could be mostly a quick cirurgical edits to make the error go away.
Edit: these are the links for the changesets on OSMCha
And yes, I just noticed the way/32137334 still with one Osmose warning (and I agree it not false positive). But these things on massive rivers is easier to edit, then wait for next Osmose update. But if the idea is start to improve a lot of rivers in a region, then might better focus on a few days).
Thanks for the long answer. Because my major interest is to update the road network in Sweden, I’m not sure if will spend any time to fix Ångermanälven. We will see.
If you are realigning rivers like this I highly recommend you use JOSM. The ImproveWayAccuracy tool makes it very quick and easy.
Thanks for the tip. I haven’t used that tool, so I shall have a look.
Is that what it’s called ;O))) Using it all the time in combo with circle arc (Shift C) and things like selecting a series of nodes on a meant-to-be smooth corner line and hit O to make it ‘round’ and even use temp dummy lines to trace a crooked bend, add nodes while drawing where e.g. a hairpin is mapped with 4 nodes but think 8-10 will make it look nice and hit O too to get a perfect 180 degrees road turn or long bend. Practice makes perfection, sort of.