Hi Jem, Am 29.10.18 um 04:08 schrieb Jem: > Re: https://www.openstreetmap.org/way/634085262 and several more like it in > the area. > > It seems that new, short ways have been introduced to replicate the purpose > of the existing barrier nodes. i.e. to prevent routing for vehicular > traffic. I believe it is incorrect and just adds complexity. > > I plan to contact the user to discuss, but want to make sure I'm right. Can > any experienced members please advise?
This is an edit of an employee or contractor of Amazon Logistics editing OSM. They documented themselves on the OSM wiki a few days ago after one email by myself and two by the DWG. (I had escalated the case to DWG because they continued editing in Germany without a documentation) https://wiki.openstreetmap.org/wiki/Amazon_Logistics My comments about the wiki page: I would like to see the driver feedback directly, not the edits only in order to be able to analyse the whether the editing errors come from the staff or if the reports are a useful source at all. It is unlikely that all the errors they fix have been reported by drivers or are derived from GPS traces. I assume that they do some automated QA checks, e.g. for routing islands. Its results should be published. This might help themselves because they can get feedback from the community to improve the rules the QA software applies. I am happy to help putting the results into OSMI or an instance of Osmoscope if they are able to give me a shape file, SpatiaLite database or anything else OGR can read. They should publish the training material in order that other companies can learn how the training should be improved. Yes, that leaks business secrets but the OSM community and their work is more important. I have observed the following issues with their edits: (1) They add very short ways to substitute barriers (e.g. https://openstreetmap.org/changeset/63524103): This should not be done. (2) They add access=no/private inside closed areas where all roads leading to the facility are access=no/private. This looks good to me. (3) They add or remove oneway=yes or change the direction of ways. It seems to be good in most cases but proper judgement is only possible with local knowledge. Are they aware of the frequent exemptions for bicycles and public transport? (4) They "fix" the same locations they had fixed in spring this year before I reverted all their edits in Europe due to too many quality issues. https://www.openstreetmap.org/changeset/58217008 I have not evaluated yet if their second attempt to fix these "errors" really hits the root cause of the issues or if it hides the root causes again. (Fixing without hitting the root cause means that they make their QA software not showing any error any more) (5) Sometimes their edits are good but in many cases they either lack local knowledge (or experience with German roads?) or the experience with the access tagging schema and many other ways how OSM works. For example, one of them recently revived a razed:highway=* by adding highway=<old value> in a construction in the centre of Karlsruhe [1]. Nearby landuse=construction polygons, a changed layout of the intersection and razed:highway=* (highway=* missing) did not keep them from editing the way and having unlimited trust in Mapillary imagery. I wonder if they trained their staff properly w.r.t. the risks of couchmapping. (6) Editing access restrictions of streets based on the visibility of cars on satellite imagery: This should not be done with satellite imagery only. It is too likely that the edit harms more than it helps because the individual cars are explicitly permitted to use the road or the imagery shows drivers violating traffic rules (but we usually map the signs). This trouble is one of the reasons why OSM should have a strong code of conduct for paid and organised editing activities protecting craftmappers and their work and ensuring that companies respect the work of volunteers. Best regards Michael [1] The construction sites of our two new tunnels in Karlsruhe make the road layout change every few days. Any aerial image and Mapillary imagery is outdated. -- Per E-Mail kommuniziere ich bevorzugt GPG-verschlüsselt. (Mailinglisten ausgenommen) I prefer GPG encryption of emails. (does not apply on mailing lists)
signature.asc
Description: OpenPGP digital signature
_______________________________________________ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk