[Data-modeling] Adopted Children

[Robert Cook]

I would support refactoring this into a CVT, even if the resulting  
instances used more graph capacity.  This would allow a single  
structure to model semantics of various levels of detail.  If the CVT  
had a "type of relationship" property, it could be set as  
"biological", "adoptive" or "step".   If this property wasn't filled  
for a particular relationship, then the semantics are coarser (the  
more general notion of a parent-child relationship that most adopted  
children understand, for instance.)  I suspect "biological" would be  
filled in only when it is well known, most likely with royalty and  
celebrities.

The main drawback of this refactoring would be that it would break  
code that depends on the existing schema.  However, we've made  
successful schema transitions in the past with ample warning.  I  
believe that can be done again.

Also, the sibling property is really a convenience -- a  
denormalization added to capture sibling relationships when parentage  
isn't known.  It's not exactly redundant, but it does create potential  
for inconsistencies.  This relationship is already modeled as a CVT,  
so refactoring isn't necessary.  If we would like to qualify the  
relationship as "step" or "adopted" (if appropriate), then we could  
add that easily.

R

On Apr 3, 2009, at 11:03 AM, Kirrily Robert wrote:

> There was an earlier discussion about doing something CVT-like to  
> allow children to be marked as natural children, adopted, step- 
> children, etc.  I'm not sure what the outcome of it was.  Robert  
> might know.
>
> K.
>
> On 03/04/2009, at 3:13 AM, Anne Semana wrote:
>
>> Hi,
>>
>> Can we have a property for adopted children be added to the /people/ 
>> person type?
>>
>> In this WP article, http://en.wikipedia.org/wiki/index.html?curid=10365878 
>>  Susan Hufford had two (adopted) daughters, Marika and Helena.
>>
>> Thanks,
>> Anne
>>
>> -- 
>> Anne Semana
>> _______________________________________________
>> Data-modeling mailing list
>> Data-modeling at freebase.com
>> http://lists.freebase.com/mailman/listinfo/data-modeling
>
> -- 
> Kirrily Robert
> Freebase Community Director
> kirrily at metaweb.com
>
>
>
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/data-modeling/attachments/20090403/df02479e/attachment.htm