[Talk-us] Facts about the world
voschix at gmail.com
Sat Apr 4 13:14:46 UTC 2015
I have been reading most of the exchange about the different approaches.
My take on it is that the entire discussion is missing something that is in
my view much more important than the quality of the data at the moment of
I am much more worried about imports for a completely different reason, and
that is data maintenance.
Even if data were 100% correct at the moment of import, they deteriorate
over time. If the data are imported directly into the main database, you
have no choice than to maintain them manually from then on, and you need to
have sufficient poeple on the ground to spot changes in the data set,
including the imported data.
If the external data were not imported, but kept in a separate
external-data layer, the situation could be different, because you would
have the option to refresh the external data form time to time. But this is
not the data model of OSM.
And in that respect there are most likely differences between countries. I
can see this difference between Italy and Germany:
Germany has many more mappers per map feature than Italy has. That is why
there are less imports (I suppose) in Germany in the first place, and that
explains why they are generally much less import-inclined than the Italian
mappers. But Germany could deal better with the data maintenance thanks to
more mappers in the field. In Italy we have many imports, but the imported
data is generally of poor quality, we simply do not have the manpower.
Formulated another way:
the quantity of imported data in an area should take into account how much
maintenance manpower is available for that area. Where there are less
active mappers, don't be tempted to compensate by by more imports, simply
keep the map simpler.
I should point out, that I have no idea what the relative mapper per
feature ratios are in he US compared with Germany.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Talk-us