Based on the Idea that the gsi imagery is more recent and the esri may have glitches I modified it to match the GSI images.

Note that I matched the road to the entering line of the esri imagery but left the internals as they were. That’s why they had a sharp turn. I did not want to make up curve data that could not be verified by arial imagery.

In some places esri is more recent so in Toyama it can be difficult to map.

I recently drove through a bypass under construction which is wrong in all imagery layers.

Thanks for keeping an eye out for this, and alerting me. Let’s continue to improve the map together.

Best regards,
Zaneo

I think the esri imagery and Japan imagery may disagree about the curve. Do we know dates for each imagery? I can correct my edits to Japan gis if that’s more current.

Why? This is a logically grouped change set.

Next time please view the bounding box with a better tool, is an equally valid response.

I see I think in this case it's my mistake. Thanks for the correction.

However I think there are a number of
cycleway:lane=shared_lane

Such as near: node/2144984480

In the case of the way you mentioned I think it's my mistake. I can change it your suggested, or please feel free to change.

I am open to suggestions!

However I think in this case if you look at, osm.org/edit?way=1058297394#map=20/35.65262/139.86360

You can see that this is a shared lane, which often has a bus interacting with parts of it.

Another nearby example: https://www.mapillary.com/app/?lat=35.656779496964&lng=139.86364231458&z=17&pKey=1057896508085679

I think that it is important to indicate that this is not a dedicated lane...

I would like to preserve that it's a shared lane, and that it is pictogram.

However I am very open to to discussion! Perhaps we could also ask Cycleosm to support these keys.

If you prefer to chat on Japan slack, or OSM discord please let me know.

Meant to be "staircase" changed to "stairs"

Was this really a "survey" ?

Thanks for updating the area! I'd be a little careful with KIBAN 2500 as it's not really up to date.

Thanks for doing bus routes!

Is it possible to get a comment on this changeset?

In the future I'll endeavor to make rivers (that I add) in Lithuania more blobby, and less squiggly. It's not a big deal to me in the end, and actually saves effort.

Thanks for being a valuable member of OSM grin. Good to see that some people can see the forest despite the trees :)

Do rivers in Lithuania fundamentally differ from rivers in the rest of the world?

Are you suggesting that any mapping done in Lithuania from aerial images is invalid if somebody has not been in Lithuania?

Have I see a river? Yes. In fact there is a famous problem called a "coastline paradox" which also applies to rivers:

https://en.wikipedia.org/wiki/Coastline_paradox

Where it turns out coastlines, and riverbanks are in fact very wibbly-wobbly when you start to actually look at them.

If usage of aerial imagery is banned without having seen the location personally. Then I think we should work together to revert a lot of the invalid data that has made its way into the dataset. There are many buildings that really look like a building from aerial imagery, but you're right it could actually car.

I cannot verify if the river runs under greenery, or if there is a surface that blocks river flow.

What would you suggest? Mapping to the start of greenery at the river edge, then adding wetlands?

Or mapping over the greenery, and assuming that the river flowed there?

Or something entirely different?

I tend to try to map to reality, not to nice rounded corners where possible.

Are you suggesting it was an automated edit that changed the tag?

Not satifsfactory, would you like to elaborate. Previously there was no river area marked. Now there is a river area marked that is true. Perhaps the river is actually larger, but the portion marked was verifiable.

I had to fix the river the centerline which was a mess and went over clear sections of land.

that's fair game.

I'm happy with man_made pipeline and intermittent. Thank you for taking the time to explain your stance.

Explaining why you reverted does not address "So first, please don't delete changesets without discussing them."

In this case the changeset didn't contain much other work, but it could have. Having a problem with part of a changeset doesn't really deserve nuking the whole changeset without waiting for a response.

Main editor of OSM is hardly a defined metric. Last time I checked the percentage breakdown varies depending on how you consider volume of changesets or volume of edits. Most automated edits are done in JOSM because iD doesn't support them.

Please take your gripes with iD up with them directly. Or that of the preset library they use.