hallo,
just wanted to check the three items i tweaked in changeset/74319960 are reasonable, as i was working blind.
thanks

hallo,
for way/702970909 you must mean lanes=1 ?

just want to verify there are no other possibilities before i repair the same quietly.

hallo again facebook,
can you look at way/437490654 that i suspect should be reversed to lanes=1?
thanks

hallo,
can you look at the only thing in this changeset, node/6781625867 , as there are a lot of invalid non-osm keys repeated by the value, so i think what you are trying to do is wrong, with nothing to indicate it is a school, thanks

hi facebook,
can you look at the `ww' key on way/703455495 that does not match anything else in this path, and identify if it meant to be something, or a mis-typed hotkey?

thanks

hallo,
v4 of way/380848798/history here added `y=yes'. i had thought this would be a potlatch2 bug cloning oneway=yes, but i see you used josm. i am suspicious as did not note a tag on the parallel opposite direction way.

can you say if this is a mistake, or if it means something?

thanks

hoi,
can you help me with the puzzle presented by node/509583261 ?

one-way seems wrong, so might you have meant a re-ordered `no' to be expanded into, oh, i dunno, no-exit, no-way-out, no-future, no-worries, or something equally negative?

thanks

yes, even the google preview in my second search result (english) summarises that.

it's not as select-and-go as the other aerials which is one reason it's not quite so well-known,

as i've been unable to sleep for days, i decided to try and analyse this without the usual visual inspection.

on the original changeset seen at changeset/74189315 which created this polygon, you see that 1868 new nodes were uploaded to osm.

when i process the node list of post-andre's polygon, i get the same number:
freebeer@Jim ~ $ grep \ 677.......$ /tmp/722581561 |wc -l
1869
freebeer@Jim ~ $ grep \ 677.......$ /tmp/722581561 |sort -u | wc -l
1868

one node is counted twice as both start and end of the closed polygon ring.

as nothing was deleted, i can compare the list of both without a diff just yet:

freebeer@Jim ~ $ grep \ 677.......$ /tmp/722581561.v2 |sort | uniq -c | grep -v -e ' 2 '
5 * 6778004510

so somehow this node was included at three extra positions in the original...
or less sinisterly, it was the start/end node and only duplicated once, your original below andre's fix:

freebeer@Jim ~ $ grep -1 6778004510 /tmp/722581561.v2

* 6778004510
* 6778004511
--
* 6778003887
* 6778004510

--

* 6778004510
* 6778004511
--
* 6778003887
* 6778004510
* 6778004510

hope that helps to understand, and see no change of significance was made.

the pools added need to be permanent. i deleted small inflatable pools seen in a different place than you mapped them. i re-shaped a large above ground pool that was visible in two separate locations that by now could be permanent and excavated. even if it stretches the osm standards of privacy and permanency, at least it could be seen to be real at least two times in the past six years or so.

the fact you had repeated problems with changing around your plumbing is, i must say, rather suspicious. most people who make such an error get it fixed in one go. someone drowning isn't going to reach for a hosepipe.

the natural=water tag does not belong on a swimming pool, as it is reserved for lakes and ponds, not backyard pools.

coming up empty-handed still for all three of these as well, which doesn't give much confidence these changes are real, and subjective undocumented ``facts'' should not appear in osm.

even today i am unable to find any links to this proposed name, no news reports, no official announcement. enclosing this alleged naming in quotes results in my local google datacentre returning a single result directly drawn from osm. it seems premature to rename this without evidence and for a major event as this i know it will be easy to fihd.

please use meaningful changeset comments to help other mappers understand what you are doing, as you have to create a new one for each edit.

being familiar with this sort of repair, i can speculate that the tool you used either put two nodes in the same spot, or included the same node twice.

seeing v2 of the nodelist takes 123 pages in my `lynx' browser, and the full history takes 242 pages, there may have been just one redundant or duplicate node removed.

as i also do this sort of repair, i don't know how to easily get a post mortem and trivially identify the now-missing node(s) or do a `diff' between versions. several dozen new cases appear daily in north america at a minimum, mostly due to an unsolved bug in iD editor, but also from josm.

nothing is posted because one or two redundant co-located points were removed. and yes, i guess it is a fine morning now (once it warms up)

do they exist? if so, maybe someone has a reason for them, like some towns are using osm to keep inventory of their trees.

if you take a look at way/416926725/history and compare the number of nodes making up the outline, it appears to have more than doubled in your last edit.

but in reality if you follow the node numbers, you will see some being repeated, often several times, as you make three loops around the building outline.

i don't know if you did anything to cause this, or whether it might be an osmand bug particularly obvious here.

i refer to my such corrections as some variant of dupe nodes, which maybe andre should consider to make it clear this is a simple geometry repair, and nothing to do with the map data, as i have seen another similar comment a day ago.

to pt53 re: potlatch2, it should be possible to use WMS backgrounds in p2 by using a redirecting url, for mappers who feel that iD is too often seen to be responsible for problem edits - pulled and glued nodes, areas frequently joined to lines when nearby, broken relations of many types, or who are just familiar with p2 as a lightweight yet capable editor while iD comes off to me as clunky and overloaded by now.

however it seems the public site does not exactly work to rewrite the url in useful form, probably a bug especially fun to trigger with `curl', as probably most users here have moved off p2 (i still see heavy use in south america and southeast asia) so use of the site is minimal.

i was hoping to provide a tested custom imagery url, like http://whoots.mapwarper.net/tms/{zoom}/{x}/{y}/webatlasde/https://owsproxy.lgl-bw.de/owsproxy/ows/WMS_Maps4BW , to be able to view and judge the aerial offset and age, but no go - in the hopes it could help anyone not feeling limited by potlatch, even potlatch1, and who does not yet want to make the leap to josm, yet does not feel comfortable with many design decisions of iD.

sorry for off-topic.

hi baden,

urk, go map!!1 again. i probably don't need your answer, but can i assume the added node and strange appearance of way/83227546/history was not your intention, but i can reasonably return it to v1?

thanks

salut,
please take a look around node/2781019856 .

it appears you have pulled several nodes and probably a section of footway out of position there.

merci