[Talk-ca] importing GeoBase Data (learning from TIGER)

[Jason Reid]

James Ewen wrote:
> Okay, 'splain this to me...
>
> Let's just say that we have the whole of Canada imported from Geobase.
> How does an update get processed? ie. Geobase makes a big update to
> it's database, and we want to update OSM from that? Will there need to
> be some ID tag on each way so that we don't just create a whole new
> layer under/over the old?
>
> If this is the case, we need to ensure that these ID tags stay with
> the ways. If someone goes in, and joins two ways that make up the
> street in front of his house, because he thinks it really shouldn't be
> two ways between the three nodes, is this going to be a problem?
>
> If we do a mass import, but skip well defined areas like Toronto, does
> this mean that it's going to need to be skipped forever since the
> Geobase tags won't be integrated into the skipped area ever?
>
> Whatever it means, I'm itching to get more data stuffed into Alberta!
> I've burnt a couple thousand dollars worth of fuel just putting some
> of the major highways into the database. I'm hoping there's going to
> be some way to define an area, and then ask for a Geobase import to be
> done on that area, and then I just need to go into that area and make
> sure things imported properly, watching for any errors/duplication.
>
> James
> VE6SRV
>
> _______________________________________________
> Talk-ca mailing list
> Talk-ca at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-ca
>   
1. Yes. Geobase assigns each node/line in its data set a unique ID for 
this purpose, so we'll just use those.
2. Yes, its going to be the single issue that may well prevent doing 
much more then updating properties of ways/nodes that have tags from a 
previous import. And even that could be full of problems. Theres a huge 
potential to cause problems even without modifying the geobase data in 
OSM, but rather adding things connected to it.
3. Yes, and No. The idea is to skip areas with data until someone 
volunteers to hand massage that data into place. That would pretty much 
involve loading both the current OSM and Geobase data for an area into 
the editor, and picking through it to remove the duplicates. Some of 
this could be automated to an extent, but its really going to be a 
manual process.

-Jason Reid