Little Brother's Diary

Useless data that only clutters the map and often geometrically wrong

Posted by Little Brother on 28 January 2019 in English.

Have a look at this:

http://tools.geofabrik.de/osmi/?view=geometry&lon=11.52821&lat=48.02831&zoom=18&overlays=self_intersection_ways,self_intersection_points

It is a sport=climbing_adventure park. Multiple contributors have mapped the path through the trees which are probably at different levels and go above/below the same path; i.e. the contributors have created crossing ways which are not acceptable in OSM unless you assign different levels to the crossing sectors.

Such errors are all over the world in theme parks and swimming chutes, etc. Armchair mappers spend a lot of time correcting errors but this error cannot be rectified. It requires local knowledge to do so.

Now I come to my grudge: Paths/tracks in theme park attractions such as water/swimming chutes, and climbing adventure parks serve no navigational purpose, are often incorrectly mapped and only clutter the map with redundant data. Rendering will also be difficult and serve no purpose at all.

Solution:

A polygon defining the site with all the relevant general attributes is generally sufficient. Roads, tracks and paths together with a single node POI defining the attraction (with any additional attribute(s)) within a site are also relevant and should be mapped.

That blade of grass is coming back to haunt us.

Posted by Little Brother on 19 July 2017 in English.

Remember that comment about the map being first finished when every blade of grass has been documented. Of course it was only meant to be figurative but some contributors do go to extraordinary extremes to map useless data.

Using OSMI I found this:

http://tools.geofabrik.de/osmi/?view=geometry&lon=6.96155&lat=51.66477&zoom=18&opacity=0.53&overlays=self_intersection_ways,self_intersection_points

When downloaded these ways are are tagged as barrier=line

They are the ground white lines of a tennis court: An entry from a contributor which has over 4,000 changesets to his/her credit. I know that there are no fixed set of attributes so what this contributor has done is within OSM rules but I think such data is redundant as they will never be rendered. But maybe this contributor is writing his own renderer.

So what’s my problem: Well, it is that the data is not being entered without mistakes which the validation applications find. If not corrected then the users who try to clean-up the data base will be continually linked to these errors which is of course a waste of resources.

Basically it comes down to the fact that many POIs can be well enough documented with a single node or area with the applicable attributes. It would then be easier for everybody.

I don’t like to delete the work of others but it give little pleasure to see redundant data like detailed water slides, etc where a single node would do the job better.