hallo amazon, how is this happening?

this changeset was uploaded three minutes after another one tracing the exact same ways through a drive-through. yet there are tagging differences between the two.

first, how can two of your mappers be working on the same thing?

second, i presume the length of time this changeset was open means it was necessary to resolve the conflicts with the older ways getting an unexpected version bump.

in this changeset, that means the upload of changeset/74160641 has become detached, with way/722414676 now being disconnected at both ends.

the corresponding way from this changeset, way/722415243 is tagged as private, which conflicts with access to the parking aisle and drive_through service drives lacking this restriction. do you maybe mean access=customers?

there is also a difference in one-way tagging; i see no obvious prohibition to leave the drive-through or parking spaces heading north.

can you please look at cleaning things up here? thanks.

also, i plan to clean up the iD-bug overlap in oregon pike, seen here at v13 (the original full-length) but in the other changeset at v2, which will probably mean losing the turn restrictions your team added at both ends.

however i am a little concerned about timing, with the time of day for these two changesets seeming to be morning in europe, nighttime in the us of a, and daytime east of europe. can you give me a general idea when amazon team members can be found editing so i can schedule my work to avoid conflicts? a timezone or utc range, plus days off, would work. thanks.

however please look into how it is possible two different mappers can pick the very same things to add within minutes of each other, and see about spreading out duties. thanks.

This changeset has been reverted fully or in part by changeset/74774305 where the changeset comment is: remove fake fiction

the idea behind the script came about seeing mass deletions elsewhere mixed with additions and other changes that would be lost by a revert, that i've had to let slide.

now i can keep an eye more closely on user plitscheplatsch (however it's properly spelled, never was enough a sandmännchen fan) when pointing out suspicious changesets, that happens periodically.

the enhancement i need to add is to work by a changeset id, rather than either a list from below (20 per page gets wearisome) or from commandline analysis of a downloaded osmchange xml file linked at the bottom of the page here.

This changeset has been reverted fully or in part by changeset/74772310 where the changeset comment is: undelete driveways that were in fact tagged as driveways and not roads

This changeset has been reverted fully or in part by changeset/74767172 where the changeset comment is: revert of mass deletions not reflected in aerial imagery

please use proper changeset comments. a single full stop conveys no info. and do not make fictional edits like this. reverted.

This changeset has been reverted fully or in part by changeset/74766670 where the changeset comment is: revert of mass deletions not reflected in aerial imagery

This changeset has been reverted fully or in part by changeset/74766617 where the changeset comment is: revert of mass deletions not reflected in aerial imagery

the use of UPPERCASE NAMES is wrong for osm, and many categories are wrong, such as bank for an insurance office, supermarket for a package store, or clothing shop for a cleaners.

what's up with the weird swimming pool north of here with the road through it? it's long out-of-date and should be deleted and replaced by the present blocks of flats.

hallo an alle,

i followed the link to osm.wiki/Denver_Planimetrics_Import to try to edumicatify meself after seeing weirdness i could not ignore.

my opinion is that rounding height to full metres is too crude. that's a granularity of half me height and could put me back in the appenzell which i blame for the repeated head injuries that has caused the brane trauma i suffer today. nice meadows though compared to california.
i'd strongly suggest adding and limiting to one significant decimal digit or a granularity of 10cm, the width of a human hand, making the difference between a faceplant and grazing me head on the beams.

then, with such accuracy, i can guess if the values given are plausible or not. my gut feeling has been there are too many houses with height 3m that do not match what i expect from the roof shape.

note i've commented on other import height irregularities elsewhere, but i expect neither you or meself to remember nor find where.

the other thing to bother me to the point of a nervous breakdown is that all the imported buildings in one area i inspected (where tagging was wrong) had on offset of about 2 metres (guessed from width of european autos; could be three metres as here may be uncomfortably close to texas, or sir guy grand [peter sellers, the magic christian, film 1969])

anyway, all the good referenced imageries (mapbox overflights, bing z20, and archived bing z19 clarity overflights) show this offset. as do newer maxar satellite nabs. therefore i'm wondering if the stated reprojection to osm's wgs-84 is somehow flawed.

note in the link directly above, the undeleted building i restored had slightly under a metre east-west offset, and maybe 20cm north-south, compared to the import, eyeballed from potlatch1 that seems to lack the more precise (E)ntfernung distance kezboard command available in potlatch2.

do i have a point in this drivel? i think not. may as well post anyway. yay me.

This changeset has been reverted fully or in part by changeset/74650030 where the changeset comment is: delete totally bogus maps.me personal castle rubbish

This changeset has been reverted fully or in part by changeset/74650008 where the changeset comment is: delete totally bogus maps.me personal castle rubbish

might help, probably won't as i don't know what-all tools are out there, but if you see my changesets of this morning they should guide you with better focus to the particular places i restored now-duplicates:

74610154
74604876

the third one here is trivial to see both buildings.

achavi will show the difference between the above and your own much denser imports.

i may as well point out i found one tagging problem and one overlap in one of your changesets, that i've mentioned elsewhere, you may also want to look at and fix (the overlap could be tricky, but josm can probably handle it and i've just thought of a hacky potlatch2 workaround)

cheers mate

nevermind; i see there's an outline from a later changeset, but lacking rendering tagging. also nextdoor two overlapping outlines, one a shed, the other not rendered, both with an unneeded node.

i'll trust your deletions were made in mentor capacity as the faulty replacement for the single item came as part of a much larger changeset. pity i didn't check out the non-building addition(s) i noticed when reviewing all changesets.

sorry...

okay, an achavi check led me to the conclusion that nrenner's copy doesn't just need a plain osm background, but also a vivid eyewatering-osm so one can actually see the difference between the pale building and the more pale residential when the former is outlined. like how osm-carto used to look years ago, only more so.

the missing building is the only apparent lacking one, along the southern row visible at https://nrenner.github.io/achavi/?changeset=69913240 .

in the upper right are further import deletions that look to all have been restored, though i've not checked for differences in tagging.

northwest in the bbox are several deletions of yer village maps contributions, where chachafish has encouraged preserving that user's history via the josm wossit function dealing with geometry that after being awake so long i simply cannot ... oh wait, replace geometry. guess i can still think, albeit with a hideous lag.

but i can't remember what else i intended to write here now.

i started at the very beginning, a really good place to start, and no further deletions from this user account were found.

what was weird was to see changeset/69754556 which was also the only gap, while picking another random deletion pointed again to this user account. OSM TURF WARS!!1! more fun than watching HOT mappers remap the same buildings five or six times each.

anyway, i'll try to figure out what was happening there. there are a small number of mapped objects not present that appear to be unimportant, maybe.

or i could sleep

mojn gela,

i'm checking another importer's work and i came across puzzling changeset changeset/69754556 which led me here.

i'm just getting started after working with chachachafishy to put back history onto remapped items, so i'll be looking more in detail sometime this afternoon if i don't fall asleep first.

just a friendly heads-up

greetz

alright, start small, but, um, there's a gap here as it seems gela had a beef.

set this one to the side for a bit

okay, this is weird that no ways are listed below as deleted, while the raw changeset seems to show 11. also 80 nodes, while 49 are enumerated below.

but wait, apparently only 1371 ways were created, yet the count below seems to include the 11 deleted.

i just have no idea on what page(s) to find them. blast. i was counting on deleted items always appearing first to quickly determine what happened.

time to tweak my skr1pt again...

chachacha, are you checking every edit like this, or have you skipped a bunch? just wondering kthxbye

they're back. man, i don't think the effort to write a script to feed the numbers below was worth the bother...