Just went through areas near me and corrected about 50 errors from keep right. How do any routing algorithms actually work when the data is so bad?
OpenStreetMap
Just went through areas near me and corrected about 50 errors from keep right. How do any routing algorithms actually work when the data is so bad?
Kaozeadenn
Evezhiadenn a-berzh wallclimber21 d'an 22 October 2009 da 01:16
They don't. :-)
Which is why it's so important to fix keepright errors!
Evezhiadenn a-berzh Gregory Williams d'an 22 October 2009 da 08:40
Remember that the data quality varies from area to area. It may be, for instance, that a mapper in a given area isn't aware of the important that connecting roads / footways etc. need to be connected via a common node, whereas mappers in other areas can be fully aware of this.
When you're discovering these errors are you sending a quick message to the person that mapped it to help educate them on how to input routable data? Otherwise that person may just continue to enter unroutable data, thus creating more work for you to fix.
Evezhiadenn a-berzh DanHomerick d'an 23 October 2009 da 00:03
In the US, long distance routing has improved a lot due to the 250 cities project:
osm.wiki/TIGER_fixup/250_cities
The project helped identify the problem areas (everywhere, at first!) and gave a quantitative track of the progress being made. For me, at least, it was a great motivator.