please help in revieing edit

Hello,

please help in reviewing these edits:
https://www.openstreetmap.org/user/txsasdx

I reverted some clearly phantasy edits. User created about three weeks ago a lot of “construction” roads. Also ref tagging was strangely modified.

I am still busy with the server upgrade, so please support here.

Stephan

I was in this area for a road-trip. Back then, the roads did not have duplicate ref numbers. I am uncertain about that.

But primary roads crossing at arbitrary points seem to be a trademark of this user, like https://www.openstreetmap.org/way/918536056/history#map=18/14.94966/103.52458

The user did more than 50 changesets. At least five of them show this vandalism pattern:

Sample:

OMG!
This is terrible. I checked one of his changesets at random #100090504 and found a secondary highway running all over the map, crossing many small roads and tracks. The user actually added bridges to cross these other highways! Maybe s/he is actually crazy?

After checking Google Streetview in a few places I started to delete it. But be careful when doing this because conflicts can result easily. It isn’t bad enough that we have people adding ridiculous data, names without any other tags, wilderness_huts all over Chiang Mai, etc., ad naseum, but now we have this crazy idiot to deal with.

It’s going to take a long time to fix this!

Each changeset I found I reverted fully. A user doing such edits isn’t doing it by accident. So I do not trust the other edits. What about the tiny manipulations of the highway refs?

User is now blocked: https://www.openstreetmap.org/user_blocks/4837

How shall we proceed? I did run into some conflicts when reverting changesets, as I did not expect to face such a big issue. Especially once objects had been touched multiple times by the user it gets tricky.

Do we need support from DWG in doing a full revert? Maybe going from newest to oldest is a good strategy on reverts? Or stacking all together and then trying to resolve remaining conflicts?

Are we certain that all edits qualify for a revert? Or is it after the trouble caused by this user not worth thinking about what to keep?

I forwarded a ping to our local mailing list. I think user also creates a mess on wikipedia. Is this a playground for some bot doing pseudo-meaningful geo-content?

https://en.wikipedia.org/wiki/User:RTX3000

Look at the history for a page created by this user (this reference was used in OSM context):
Within minutes, it rotates through random geographical areas. It smells like bot.
https://en.wikipedia.org/w/index.php?title=Draft:Asia_highway_120&action=history

Hard times coming, if people start firing with manipulation bots at our data.

PLEASE DO NOT manually fix the data. Do full reverts of changesets.

I’m interested in helping but I’ve never done a revert of a changeset before so I’m a bit unsure of the best way to help. Yesterday, in an effort to help, I had intended to merely delete the added highways in the changeset I mentioned earlier but when it resulted in conflicts I abandoned my edits and quit. The conflicts were my fault because I hadn’t downloaded the entire fake highway before deleting it.

I wonder how many other users are people like this? It appears to be pure vandalism yet why would this vandal take the time to add bridges to his fake highway?

I think in the interest of saving time, doing full reverts is the best way forward. If the user has added actual real data that will get lost. So it goes.

I’ve just reverted all his changeset fully.
Btw, some of them have had edited manually before, so it won’t effected.

Thank you for reverting!

I had checked before the first changesets of this user and all of them bad bogus data.
The block if this user will expire in a few days. I wonder how to set an alert in case the edits continue.

Out of interest: Did you manually loaded each of the fifty changesets into JOSM reverter plugin, or did you have an optimized way? I used reverter on some of the earlier changesets until I realized the full extent of this mess.
Also, as data was touched, it resulted in a larger number of conflicts. Hope nothing too severe broke on the resolution of these.

This one we spotted quite early after three weeks, other edits might go unnoticed for longer.
I can just ask for more reviews of edits, like on OsmCha to spot such behavior early.

I manually reverted by a JOSM reverter plugin stacking one by one, from latest to the first edit. I think this is the most efficient way and lead to no problem except for geometries which was edited after all changeset, by another user.