[Data-modeling] influence nodes are not always people

[Dan Milbrath]

I guess I question the practicality of taking everything down to the people-centric level.

Is it useful to know that the Orange revolution was influenced by the Rose revolution? Is that a question you might want to ask without knowing the people involved? Same holds for musical groups. I think so. 

I guess if people here are generally are opposed to this, we could model a new property on musical group that allows us to make this kind of assertion (influenced/influenced by), but its a shame because at that point we're unable to get all of a band's influences via a single property query - we'll likely resort to only showing what bands influenced what other bands -- and miss the fact that a band may have been influenced by an individual not in music, a movement, en event, a city, etc...




----- Original Message -----
From: "Iain Sproat" <iainsproat at gmail.com>
To: "Freebase data modeling mailing list" <data-modeling at freebase.com>
Sent: Saturday, December 13, 2008 2:26:59 AM GMT -08:00 US/Canada Pacific
Subject: Re: [Data-modeling] influence nodes are not always people



-1 
I think person should remain as an included type. 
Inanimate objects and events cannot be influenced, only the people who create or participate in them. 


In your example it is actually Andy Warhol influenced Lou Reed et al of The Velvet Underground, who then presumably influenced Anton Fier et al of The Feelies. 


As another example Ukraine's Orange revolution is often stated as being influenced by Georgia's Rose revolution , but in reality it is Viktor Yushenko and his followers who were influenced by the likes of Givi Targamadze and others. 


It no doubt makes querying slightly more complicated, but taking the people-centric view correctly attributes influences to the people involved. 


On Sat, Dec 13, 2008 at 6:12 AM, Dan Milbrath < dmilbrath at metaweb.com > wrote: 


see: http://www.freebase.com/type/schema/influence/influence_node 

It seems like we've gone to some length name the influence node type in a pretty general way, so it strikes me as odd that we've asserted that people/person is an included type on it. 

Why not use influence node on other things like musical artists, political events or art movements. For instance, I might want to assert that Andy Warhol influenced Velvet Underground who influenced the Feelies. 

What do people think about removing person as an included type? 

_______________________________________________ 
Data-modeling mailing list 
Data-modeling at freebase.com 
http://lists.freebase.com/mailman/listinfo/data-modeling 


_______________________________________________
Data-modeling mailing list
Data-modeling at freebase.com
http://lists.freebase.com/mailman/listinfo/data-modeling