[OSM-talk] iD invents nosquare=yes for buildings which should not be squared

Pierre Béland pierzenh at yahoo.fr
Fri May 10 19:05:05 UTC 2019


May 20 2019 at 14 h 02 min 51 s UTC−4, Stefan Keller <sfkeller at gmail.com> wrote : 
> Trying to get focus back on the thread topic.

> Storing hints like nosquare=yes (or square=no) is not best practice of
> data curation on w worldwide level.
I dont think either that this is the solution.  We have to look where these problems come and try to correct from the source.  Adding  such tag will not help software tools to identify where we have problems and can create some missunderstanding about its meaning. And experienced mappers are more and more reluctant to correct inprecise building mapping.

For Newbies, Building Quality Analysis last year this show some Tasking Manager projects with some 60% of unsquarred buldings (see https://opendatalabrdc.github.io/Blog/#!Database_Quality_Analysis_Tasking_Manager.md). 
The same problem for the DR Congo Ebola response in november where we had to restart mapping Butembo in an emergency response and restrict mapping to experienced mappers.

For an editor like iD, it can participate with other solutions like better training to assure that Newbies understand what Building tracing really mean and give then some feedback, like for example saying before save "You have 10 over 12 buildings that are unsquarred.  If you dont know how to make rectangular buildings, you should ask for help before you continue.  Do you want to send data anyway ?"

We have the same problem with some Building import projects and I think that this needs to be fixed where it originates, before it goes in the database.  For newbies, this mean more training and monitoring in particular in the context of Mapathons.
For Imports, this mean to analyze carefully and correct the data before the Import.
 
Pierre 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20190510/aa781ce2/attachment.html>


More information about the talk mailing list