On Mon, 2016-06-20 at 11:26 +0200, Martin Koppenhoefer wrote:
> 
> 2016-06-19 22:35 GMT+02:00 Ilya Zverev <i...@zverev.info>:
> > As for the maps.me, I am glad that foreign names issue is basically
> > the only one that most people agree on.
> > 
> 
> there are lots of different issues, and even if many of them have not
> yet commented on them, I still believe they do have the potential to
> harm overall data quality. From the manual reviews I have performed
> so far, the amount of new issues introduced was far bigger than the
> useful information that has been added, but I didn't look at enough
> data to make this representative in any way (of course).
> 
> Some of the issues that come to mind:
> 1. stuff put projected to the middle of the road rather than the
> actual position (common newbie error, possibly because that's how
> google and others present search results)
> 2. stuff put without a tag what it is (just a name and a property
> like tourism=attraction)
> 3. duplicates added (things that are already there)
> 4. poor semantic level (very low detail in tagging, in some
> occassions to a point where it becomes not understandable any more,
> sometimes mistagged as something vaguely similar)
> 5. sometimes missplaced objects far off (likely due to bad location
> data in the device and users not familiar with the area, and not
> willing to properly orient themselves)
> 
Many of these are newbie errors, the same as we see with any other
editor however the big difference I see with maps.me is the apparent
lack of reaction to changeset comments that allow the community to help
newbies through their initial edits. When adding comments to maps.me
changesets I do get the feeling I am wasting my time.
Another observation I see is the geographical spread of edits, a bar in
Portugal followed by a guest house in the UK about the local knowledge
of what they are adding.
Phil (trigpoint)

_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Reply via email to