[OSM-dev] Proposal for OpenMetaMap - proper OSM import solution
jaak.laineste at gmail.com
Thu Aug 18 10:41:30 BST 2011
2011/8/16 Mike N <niceman at att.net>:
> On 8/15/2011 4:57 PM, Jaak Laineste wrote:
>> But it is not plain "banning imports", it is to provide alternative
>> tool (I call it OpenMetaMap) which enables to link various external
>> data sources with OSM data, with much better flexibility than with
>> current import approach.
> Also it might require that map editors deal with complexity:
> 1. Take a survey
> 2. Add a "museum POI" from the survey. Now there is an external museum
> POI source to be aware of. Is mine a duplicate? Should I add to OMM
> instead of OSM? How to learn of all these other types of items that may
> exist in the area just surveyed?
Right, there are many complex cases. In this case I can see following.
3. After survey you open map with JSOM. It opens/loads two layers: OSM
data and external sources for the area. You should see now that the
museum is already on the map - just it is from the external source. In
my survey I found some additional tags and maybe more accurate
location - so I just edit the point and save it. It would look like
just like with any other existing POI.
4. Under the hoods JOSM creates OSM object with my tags and OMM link
to external database.
This was quite simple. More complex is when I do not add any new data
compared to the external source, just confirm it. Then I have to make
a) I am happy that the object is already "in the cloud" and on the map
- I do nothing.
b) I want to have it also in OSM. Then there is special JOSM action
for objects "internalize object" which means creation of object copy
in OSM and adding OMM Link for it. Maybe even should have advanced
option to select which tags should I copy - I should pick only the
ones I really can confirm based on my survey.
> This might not be appropriate for all imports. Everyone likes to point to
> problems in the TIGER dataset, but in areas where it has been adopted by the
> community, it has become a first class dataset in its own right. It would
> not work as an OMM dataset in my opinion, except serving as an alternate
> source of data for adding to the primary OSM map.
That's true, it is not a magic bullet. If OSM community is much
better than external database maintainer then OMM can be overkill. In
many cases the external database is dead anyway.
More information about the dev