#20674 closed defect (fixed)
Water area inside water area not detected for waterway=riverbank
Reported by: | Owned by: | team | |
---|---|---|---|
Priority: | normal | Milestone: | 21.03 |
Component: | Core validator | Version: | |
Keywords: | Cc: | skyper |
Description
When one waterway=riverbank
objects is inside another, no warning is shown in the validator.
However, when two natural=water
objects overlap in this way, the "water area inside water area" warning is reported.
Test case demonstrating the issue attached.
Attachments (2)
Change History (9)
by , 4 years ago
Attachment: | overlapping_riverbank.osm added |
---|
comment:1 by , 4 years ago
I thought about adding this test when I fixed #20570. These tests are rather slow, so the question is if they are useful. Did you see this problem in real world data?
comment:2 by , 4 years ago
Yes, I've encountered this in real world data, though I've fixed it each time I found it, so I don't have a ready example to show. The issue is that the value "riverbank" sometimes makes people think it is for tagging the sides of rivers (rather than water-covered areas of rivers) and as a result of that, sometimes you'll find river islands tagged with "riverbank", with people thinking they're tagging the sides of an inner riverbank. But really, they're just making the river island invisible :) The only way I'm even able to detect this error is to convert all riverbanks in an area to natural=water
+ water=river
, and then the validator works properly and detects these cases.
This misunderstanding of the value "riverbank" is why I've reported #20591 Rename "Riverbank" preset to "River area", to help reduce this confusion.
This situation/confusion is yet another reason why waterway=riverbank
should be replaced with natural=water
+ water=river
and I've been working at a grassroots level with local communities to fix river issues which includes upgrading this tag (see https://wiki.openstreetmap.org/wiki/User:ZeLonewolf/Procedure/River_modernization where a number of mappers have been collaborating).
Personally, I would like to see JOSM recommend auto-upgrade of this tag, just like iD has done since late 2018 (https://github.com/openstreetmap/iD/issues/5591), but I recognize (see: #17874, #8759) that not everyone seems to agree.
In the absence of JOSM recommending upgrade of riverbank, let's at least fix tagging issues associated with riverbank-style tagging on an equal footing with those of natural=water
.
by , 4 years ago
Attachment: | 20674.patch added |
---|
comment:3 by , 4 years ago
Cc: | added |
---|
@skyper: I just wonder why we have the check for :areaStyle in these tests. Do you know why?
comment:4 by , 4 years ago
No, I never took care about :areaStyle
. I'd say this is not needed for positive search/regex and its origin might be the general rule (L206):
area:closed:areaStyle ⧉ area:closed:areaStyle { throwOther: tr("Overlapping Areas"); }
The two rules (L191) above probably do not need it, but below (L216), I guess, it is needed as not all closed ways with natural=*
are areas.
/* #9556, #20570 */ area:closed:areaStyle[tag("natural") = parent_tag("natural")] ⧉ area:closed:areaStyle[natural][natural !~ /^(water|wetland|coastline)$/] { throwWarning: tr("Overlapping Identical Natural Areas"); }
Just, my 2 cents.
comment:7 by , 4 years ago
Milestone: | → 21.03 |
---|
Test case demonstrating the problem.