[Data-modeling] Proposed changed to/aviation/airline_airport_presence

[Jeff Prucher]

I think the property could just be deleted. If it's just meant to capture
the cities near the airport, it's redundant. If it's meant to capture actual
routes, it's insufficient. That is to say, since the property isn't
reciprocated, destinations need to be entered twice, once from each airport
(e.g. SFO -> JFK and JFK -> SFO would be two CVTs the way it's modeled now).
However, because the property is non-unique, just reciprocating it wouldn't
work either, because the other destination airports would show up as well.
If we want to model airline routes, we should look into a new property.
(Although this data changes all the time, so I'm not sure how useful it
would be to capture without some way to keep it reasonably up-to-date.)

Jeff

> -----Original Message-----
> From: data-modeling-bounces at freebase.com 
> [mailto:data-modeling-bounces at freebase.com] On Behalf Of Bryan Cheung
> Sent: Wednesday, August 13, 2008 2:59 PM
> To: Freebase data modeling mailing list
> Subject: [Data-modeling] Proposed changed 
> to/aviation/airline_airport_presence
> 
> The property cities_served on type /aviation/airline_airport_presence
> is a bit confusing. Is this disambiguating, non-reciporated 
> property trying to describe the cities that the airline flies 
> to from this airport? Or the city of the airport, which is 
> then redundant with the serves property on airport?  This 
> ambiguity also leads to some incorrect type assertions as 
> some users add the airport as the destination city.  Perhaps 
> the property should be destination airports, rather than 
> cities.  What do people think?  The expected type of this 
> property would be updated to /aviation/airport, and the 
> property name would be updated to something along the lines 
> of airports_served .
> 
> Bryan
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling
>