For start: sorry for accidentally running a bot edit outside Poland that was discussed only with Polish community. Edits were limited to adding/editing website tags. Main problem was that bot was run without required discussion, OSM data should not be damaged.
In Germany it resulted in adding or improving 1105 website tags (also, 1400 more across world - also in locations where bot edit was not cleared to run at all).
Root reason was that location was supposed to be limited to a Poland but required check was not actually performed in the final version of code.
Luckily this defect should be independent from data processing, so website tags should be fine and not damaged.
Edits themselves should be not be breaking data and from spot checking they were fine as far as values of website tags are concerned. But edit was not discussed so it should not have happened. I had nice plan to first run more limited edits, figure and fix problems if any and then propose run in areas that are less bot-happy like Germany to make more likely they will be accepted.
I guess that this plan is thoroughly bungled now.
I am thinking right now what I should do with it, probably I will revert these undiscussed bot edits. But it will take some time, as I need to figure how to do it properly without causing any further issues.
Mir war auch eine Änderung davon in meiner Nähe aufgefallen. Diese hatte ich aber überprüft und sie war richtig. Von daher: Bevor Du die Änderungen jetzt wieder rückgängig machst, kann man sie sich doch einmal ansehen und dann entscheiden, ob man sie behält.
I had also noticed a edit in my neighbourhood. But I had checked it and it was correct. So before you revert the edits now, you can have a look at them and then decide whether to keep them.
The tricky issue here is that edits (new website values) should be better than previous one, and as far as I checked are.
This part was quite well tested and is quite careful about editing.
So revert would be more on procedural grounds and to make more likely that people will agree once actual bot proposal is made. Not to fix any data damage - revert will make website tags worse.
Infolge der jüngsten Entwicklungen, siehe Fehlgeschlagene BOT-Aktion? "ATP-Import" Hä? ziehe ich hiermit meine Meinung zurück und fordere einen vollständigen Redact(!!) der Bot-Operation…
??? maybe try writing it in German and let people to read it directly if they know German and let other people to translate with autotranslate
As far as I know redaction is not a correct tool here either way.
Data redaction (distinct from reverting) is reserved for copyright violations, people posting highly problematic content like private personal data. Not for underdiscussed imports where data is not a copyright violation.
(also, vast majority of my bot editing was independently approved in a correct order, messing up this edit does not retroactively invalidate other permissions)
Es ist mittlerweile alles gesagt worden und es ist auch kein weiterer produktiver Erkenntnisgewinn zu erwarten.
Es tut jedem gut, in einem Streit einen Schritt zurück zu treten und sich zu fragen, ob das, was man vertritt, sinnvoll und im Projekt mehrheitsfähig ist oder ob man sich verrennt hat.
