Automated edits/Ben d8a2ae3a bot
What
Change dead links from e.g. website=https://dead.link/ to disused:website=https://dead.link/.
Why
The OSM-data contain masses of dead links of various kinds:
- offline: https://edeka-kusnezow-lewandowski.de/ https://www.biergarten-zum-saliter.de/
- broken, not even 500: https://www.altesbrauhaushemmendorf.de/ https://www.dezbuehne.de/
- 500: https://www.hotelamkamp.de/ https://www.pizzafun-ol.de/ https://innowoprint.de/
- taken over by domain hoarders: http://aulendorfer-hof.de/diner.html http://www.optik-millies.de/ http://www.quedlinburg-pension.de/ http://www.minigolf-hagnau.de/ http://praxis-dr-siebling.de/
This makes it look unreliable, outdated, and much less useful.
Let's fix that!
Note that simply *removing* those tags would be a bad idea, as an outdated website is an easily-detectable marker to find possibly-outdated nodes/amenities/ways/etc.
Where
Germany. I'm German, so I find it trivially easy to inspect and evaluate lots of German-speaking websites.
The code is open-source, so feel free to run this on your own country/area/whatever.
How
- An extractor downloads and analyzes the German map, and extracts all URLs from a set of tags, like
website,contact:website,contact:url,source:url,website:booking,internet, etc. - A crawler visits each domain once a month. By "domain" I use the interpretation "registered domain as defined by the Public Suffix List". A URL inside that domain is picked at random, with preference for URLs that have never been visited before.
- If all requests to a domain fail for the exact same reason for 6+ months in a row, that domain is considered to be dead.
- (proposed) Every now and then, pick a geohash area with sufficiently many dead links, and upload a changeset to add the
disused:lifecycle prefix.
Consultation
The consultation process started on 2024-01-24 at [1].
Hence, the bot is not creating any changesets yet.
When it starts, I will start slowly, to give ample time for problems to be detected. In the beginning, I will probably double-check everything manually.
Result on 2024-02-02: No automated edits for now, let's try how well exporting to Osmose works in practice. (I expect this to take at least until 2025.)
Concerns mentioned in the AECoC
- This bot is in the category “Useful edits that would be tedious to do manually, only after approval by the community and appropriate discussion.”
- None of the “Problematic usage”s apply.
- [2] TODO
Who
TODO
Logs
TODO