[Talk-ca] Conflicts between GeoBase and OSM attributes

[Richard Degelder]

There has been some discussion about the potential of conflicts between
the OSM map and when we do any updates concerning the modification of
some of the attributes assigned to the OSM way.  But, to me, there is no
need to consider the potential of an edit war between OSM mappers and
scripts used to import or update data from external sources.

With the initial import of the GeoBase data there is going to be a great
deal of work by individuals in fine tuning the import to resolve
conflicts and tune attributes to better reflect reality.  Also part of
the initial import will be assign the appropriate GeoBase NID to the
ways within OSM.  Once that is complete we have done the majority of the
work for any updates coming from GeoBase.

When we get an update from GeoBase and want to import that update into
OSM our update scripts should first check for all ways with an NID and
eliminate them from the update unless the GeoBase update has changed
them for some reason, such as created a new intersection or shortened
the road.  This will eliminate the majority of the ways within the area.
What we are not going to worry about are ways that have been modified by
individual users within OSM.

So if a user adjusts the way slightly or changes some of the attributes
for that way we are going to leave it alone as long as it has an NID.

An update will only worry about the ways within OSM that do not have an
NID and those that GeoBase says have changed in some way as well as new
streets within the GeoBase update.  These will have to go through a
system very similar to the initial import and, if matched to a new road
within GeoBase, will receive an NID as a result.

If we want to add new data, such as street names and/or addresses, we
are going to look at all of the NIDs and add new attributes as
determined from the GeoBase data based on the NID that is already there.

Basically if a roadway has an NID then from that point on the attributes
will only be modified by individual OSM users and not from a script.  So
when James Ewen mentions that according to GeoBase, and our conversion
scripts, Highway 2 by Edmonton has a Trunk status but someone has marked
it as a Motorway within OSM then with an update from GeoBase the update
script would look for the NID tag and, if present, ignore the
discrepancy in the type of roadway and not change it.  This should, at
least after the initial import, eliminate most cases of an editing war
between the users within OSM and the update scripts.  This would also
reduce the possibility of additional attributes, such as the speed limit
and the number of lanes, being lost as well.

Richard Degelder