moin, doktor bacon...

i'm trying to figure what went wrong with way/48012847 which it appears this changeset recycled the southern lake nodes to form part of a residential boundary.

do you think you could figure how to fix things in josm, as the v3 shown looks unchanged since it looked good, and only the nodes have been altered individually, which will be a pain without josm to repair node by node.

thanks

This changeset has been reverted fully or in part by changeset/72728222 where the changeset comment is: misplaced church spam as usual

iggujja, gseethar, and anyone else who may be monitoring this discussion, sorry for the late follow-up here, but...

in the case of barriers, temporary or permanent, the way of drawing the road in the esri imagery state, is not to leave a gap where a barrier is suspected, but to close and connect the ways to reflect what is actually constructed.

then the barrier is added as a node as part of the way, as presumably the visible barriers do not apply to all modes of transport such as pedestrians and likely cyclists.

only if there is an obvious gap between the roads should they be unconnected (and usually there will be something to connect the two, whether a footpath, a short section of construction, or similar).

this would also be the case if there were to be a permanent gate erected at this point to limit the traffic.

the difficulty there would be to determine the proper access values to be used, for which an on-site inspection or recent street-level photos should be used.

with the gaps as mapped now, routers will avoid using these segments, regardless of their present status, diverting people who could give feedback from actually checking out the situation unless they wonder why they were directed a less direct inefficient route.

it seems that you are using the wrong canned follow-up form-letter here, as i can see no internal terminology in use (and then even when true, i don't see a real need to spam the changeset comments pointing out their use so long as the description gives an accurate description of the change).

in this case it seems you simply did not update the description and re-used the description applicable to changeset/72453145 , which happens often enough by mistake.

no worries.

hallo georg,

sorry that i commented on the notes you opened before i was aware of this changeset discussion.

i would update the map to reflect the maxar imagery as it is the newest and probably itself some months old, so that some of the house outlines in the earth seen in maxar, may very well be completed as structures now.

i feel it is better to have the map reflect the completed state of the aerial imageries of a completely connected neighbourhood, than to await on-the-ground confirmation of the same, which in my experience, may not happen for years in this part of the world.

of course, there is the danger someone will report later that there is no access as posted by signs, as i am well aware happens often in comments to remote mapping in the uk, and i suspect gseethar has experienced this far too often, where there actually are on-the-ground mappers to offer feedback, that i expect to be unlikely here.

schoenen abend noch

mojn,

i have already repaired the errors i found, so you do not need to revert anything.

you can see what i have done, explained in my changeset summaries to way/704828410/history . this is similar to the issue i raised in changeset/72255504 (i have not viewed the fix for that yet).

your nearby way way/704828407 and connecting ways, also added in this changeset, did not have the self-intersection problem, being more complex.

as far as the username, it was obvious you map for amazon, but i did not know the history, as there are now too many mappers with different sponsors that i can no longer connect a username with their sponsor, without the help of the suggestion i made in the changeset i linked above.

thanks for adding the text to your new profile. i do not understand what you mean by ``issues'' with your old username that i can see is not deleted with 467 edits at @maakra (missing the default picture, whatever that indicates), but i hope you know you can change the username and your new username will still contain your edit history.

so, if you wish to indicate the maakra account is no longer in active use, you can change that name to something like maakra_inactive - depending on company policy mentioned in the other changeset comments, that i have not seen explained further.

in any case, as the self-intersection error that i repaired here has been frequent in the past days, i hope that all amazon mappers have now been trained to avoid this. i will soon see, as i resume repairing these errors that i wanted to do last night (i slept instead after a night without sleep)

cheers

This changeset has been reverted fully or in part by changeset/72351017 where the changeset comment is: this is not the real name, please keep fiction out of osm

can you check what you have done at https://nrenner.github.io/achavi/?changeset=72124018 ?

it seems you have not just deleted polygons, but you have somehow managed to mangle others in a way i cannot comprehend, effectively scalping them and causing them to cross themselves, throwing up errors in osmi and failing to render.

there is a large number of these errors for georgia and i have just examined a few, but is this what you intended to do, or is iD somehow deciding what to do instead?

the error that came to my attention does in fact contain mostly trees in imagery, even if approximate, so i cannot tell if your plan was to delete any imprecise polygons or if it was limited to the three deletions in bright red of ways listed below.

you may do better to see the dark red of your geometry changes by changing the achavi base layer to something like osm pale, or osm grey, or lose the new green against plain osm.

thanks

hi,
are you a part of an editing team? i can only find hints that make me 99% certain you are, after i repaired one of the service drives added here and following the example of a common error (two actually) these days.

thanks

hi,

your edit here has been flagged as self-intersecting and overlapping by the osm inspector.

i have been spending all night repairing many of these from unknown users, but this one caught my eye as having no intersection nodes whatsoever and a great example of how not to map.

i see you are using the iD editor while making a large number of edits. may i suggest learning and switching to josm, which i hope would make it less easy to upload re-traced tracks like this, or other errors it seems iD introduces too often?

i am leaving this unedited for now as an example for learning. in way/704484191 at v1 you can see how nodes as listed ...2079 and 2083 appear in that order, then soon after appear again in reverse order.

node ...2077 which terminates the way also appears a second time.

as part of your training, it should be made clear that re-tracing over part of the same way should be avoided, otherwise it will be flagged by
https://tools.geofabrik.de/osmi/?view=geometry&lon=-88.21127&lat=41.81137&zoom=17&baselayer=PublicTransport&opacity=0.66&overlays=self_intersection_ways,self_intersection_points,single_node_in_way,duplicate_node_in_way
twice here.

also, may i suggest adding your mapping sponsor to your username, so when viewing histories from my editor it will be immediately obvious that apparently unrelated users making the same mistake, are in fact related and a changeset comment like this is merited.

in an earlier error, i immediately associated one problem edit as accidental by a lyft editor due to use of name_lyft and unrelated to the purpose of the entire edit. it would be nice if i am correcting systematic errors by amazon logistics of self-intersecting driveways and service drives, that be obvious simply from the history, that these are not simple random beginner mistakes as i have been assuming.

thanks

servus,

can i ask you to take a close look at the southeast end of this changeset at https://nrenner.github.io/achavi/?changeset=72181267 ?

it looks like something has gone horribly wrong down there, with self-intersections and overlaps and i don't know what-all, as i view the raw data pointed out by osmi. can't make out if something was grabbed and pulled yet.

merci.

moin moin,

it looks like you managed to grab a sidewalk in this changeset, pulling it out of position and bending maybe a dozen attached ways, as can be seen at the very top of this at https://nrenner.github.io/achavi/?changeset=71987010 .

is this something you can easily fix in josm?

thanks

isn't it worth noting that nearly all the changes here have to do with houses to which are (or have been) added personal family names in violation of osm policy (such as [Casa ]Familia Garcia Tequia), names that should not be added to the map?

while it can be seen from aerial imageries that it obviously exists.

and it can be used for deliveries, paper routes, and visitors even if irrelevant to the general public other than being a landmark for orientation.

any good reason not to add it? access can be tuned to fit, but as a dead end, i see no reason for a router to consider it.

while it can be seen from aerial imageries that it obviously exists.

and it can be used for deliveries, paper routes, and visitors even if irrelevant to the general public other than being a landmark for orientation.

any good reason not to add it? access can be tuned to fit, but as a dead end, i see no reason for a router to consider it.

is it possible that andrew was referring to the nearby way/207068104 , just southwest, and matching the decade-old clarity imagery in use the years ago it was created?

not touched by thus changeset at all, but no longer matching current imageries.

that was touched this year by a user dhaird i do not know, allegedly according to mapbox.

all this user's additions look implausible to pokemon through-houses vandalism, but i am too weary to revert the lot to match the aerials.
sigh.

i think i know what you are seeing -- my potlatch2 editor shows the old layout at z19, but when i step to z20, i get no tiles.

https://a.tile.openstreetmap.org/19/91147/183005.png is the rendered data for part of the building, which maps to
https://clarity.maptiles.arcgis.com/arcgis/rest/services/World_Imagery/MapServer/tile/19/183005/91147
which is
‘91147’ saved [13587/13587]
91147 is a 256x256 JPEG image, [...]

at zoomlevel 20, the cycle map url could be
https://c.tile.thunderforest.com/cycle/20/182293/366010.png?apikey=6170aad10dfd42a38d4d8c709a536f38 today.

feeding that to clarity gives a permanent redirect rather than serving the tile from clarity proper:

https://clarity.maptiles.arcgis.com/arcgis/rest/services/World_Imagery/MapServer/tile/20/366010/182293

Connecting to clarity.maptiles.arcgis.com (clarity.maptiles.arcgis.com)|2600:9000:20ad:3600:15:6f24:c5c0:93a1|:443... connected.
HTTP request sent, awaiting response... 301 Moved Permanently
Location: https://wayback.maptiles.arcgis.com/arcgis/rest/services/World_Imagery/MapServer/tile/4383/20/366010/182293 [following]

as this is an ipv4 redirect, my network config prevented me from following it until now, which gives me a further redirect to the us west coast and the jpeg you describe.

why the wayback archive would mirror current imagery, i cannot say, but strictly speaking, this should not be considered part of the clarity layer proper.

thanks for the example.

i'll probably go back to disabling that misleading redirect for me...

serwoas bathis,

i am guessing you are part of mapping group given the flood of commentary of late, although it is neither obvious, nor relevant.

in any case, not seeing the unmodified entry of this changeset, i'd like to propose that you expand your aerial imageries to include the archive and sometimes leaf-off imagery of old bing as esri clarity. and in this area which blindly i am guessing to be in part of florida from past mapping altercations, the current esri imagery looks to be good if from a lighting angle that i cannot parse with my present latitude,

horza, i thought you had no seasons which is reason for me to call your haunts me home. on the other hand, that means leaves, which means lousy aerial mapping.

arrgh.

cheers. and ignore me. please.

hola dazz-jazz,

not sure if it is still relevant, but mebbe a month ago there was at least one such amusement ride outside the us but close enough to appear like a suppurating sore, in toronto or something.

i've self-aborted upon encountering any rollercoaster overnight when i can't sleep and eagerly await the osmi update to thwart your repairs, in revenge for visiting places you repaired an hour or so earlier, you b*st'*d cheating wake-then time-zone no-goodnik. oops sorry for insulting a fellow mopper.

anyway if you could purge those ferrin' entities i'd love it, if you have the experience to do so without leaving the dog's breakfast i'd be placing there.

as far as the shared nodes go, i've a great problem with the wiki suggestion to join roads with roof building parts. also where a footway crosses a road but legally, neither traffic may diverge from its own path, the latter being a case of over-thinking concepts as i tend to do when awake at night and trying to map errors to put me to sleep, futilely.

and re-reading this, it makes no sense, so me dementia is worse than expected. as usual. gaaah.

ignore me.