[OSM-talk] DASFIF airport data

Simon Hewison simon at zymurgy.org
Fri Apr 20 11:51:42 BST 2007


Frederik Ramm wrote:
> Jon,
> 
>> I wrote a script to convert DAFIF airport and runwat data to OSM but
>> I've never uploaded it to SVN since no one has provided a clear  
>> 'ok' as
>> to the legal status of using the data.

I think that if you've got a DAFIF file pre-October 2006, then that is 
regarded by most as public domain (as in US taxpayer funded).

> I have the impressing that waiting for anyone's ok (or even waiting  
> for votes) doesn't get you far in this project. I'd say float your  
> idea on the list (as you have done, I believe) and if nobody cries  
> foul then just do it. Worst case, you can still remove things again  
> if you put in a proper source tag.

Agreed.

For the UK, the DAFIF data doesn't really cover the small airfields in 
anywhere near as much detail as the UK AIP - which is, so far as I am aware, 
is copyright NATS Limited, who explicitly say no derivative works; so I reckon 
the commercial flight guides like Pooleys probably have a licensing agreement 
with NATS to create derivative works.

Because it is nowhere near a complete as the "official" versions for the UK, 
I'd suggest that it's actually been compiled by the US Government.

> The only reservation I'd have is that you have to take care not to  
> duplicate airports that are already in the database. You'd either  
> have to see to that manually, or maybe use pieces of Joerg Ostertag's  
> osmfilter script (which automatically converts GPS traces to segments  
> and checks the server for possible collisions).

Agreed. You don't want to go and stomp over airports already there, and as 
such, something like osmfilter would be the way to do that.

> Of course if your DAFIF data is more precise and/or more detailed  
> than what is already there, you could also just delete the existing  
> stuff.

You'd want to manually check each airport for which there are possible 
conflicts with existing data to see which one is more complete, and manually 
de-duplicate the data that looks apparently dirtier. Bear in mind that for 
some airfields, there's already taxiway data that isn't in the DAFIF data - 
for instance Elstree airfield - EGTR, and those taxiways join up to already 
pre-existing (traced from yahoo imagery etc) runways.

-- 
Simon Hewison




More information about the talk mailing list