[Data-modeling] I would like to correct an error in the Comet schema (Astronomy Domain)

[brendan neutra]

An aside, this is a really nice property of, er, freebase properties.  For 
example, a person who is asserted to be a "Discoverer" of a comet may not, 
institutionally or otherwise, be considered an astronomer (may not even be a 
person, thanks fluffy!). The act of creating the link in the freebase.com UI 
does assert the "astronomer" typing on the target object but the assertion is 
informal, you can easily remove that assertion without wiping out the link: 
nice. It's almost like a hint, except it makes you, at least initially, take the 
hint!

That said, I think asserting that a discoverer is an astronomer (and therefore a 
person, as well) is going to be right 99% of the time. It passes some unspoken 
utility test...

Brendan

Alec Flett wrote:
> Hey gordon -
> you don't have to remove the relationships if you just cotype the 
> persons as astronomers, and change the expected type..the links between 
> topics don't have any knowledge about expected types on either end.
> 
> Alec
> 
> Gordon Mackenzie wrote:
>> The type Comet (unlike Planet, Minor Planet, Dwarf Planet, Asteroid)  
>> the link for Discoverer is directly to type Person instead of the more  
>> semantically desirable Astronomer (so we can easily list Comets  
>> Discovered on Astronomer/Person topics).
>>
>> I would like to conform this schema to the other Astronomy object  
>> types. I will need to remove about 29 relationships and then re-add  
>> the discoverers (via list pusher).
>>
>> Will there be any impact to anybody's mashups/apps?
>>
>> ~ Gordon
>>
>> <<< gordon at metaweb.com >>>
>>
>>
>>
>> _______________________________________________
>> 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