[Freebase-discuss] Schema for deprecating properties and types

Jeff Prucher jprucher at google.com
Fri Nov 4 16:32:07 UTC 2011


On Thu, Nov 3, 2011 at 6:08 PM, Paul Houle <paul at ontology2.com> wrote:
>  On 11/3/2011 8:22 PM, Thad Guidry wrote:
>>
>> +1 Haha... we should have done this 2 years ago !

Yeah, no kidding. Ah well.

>> Isn't it just easier to do something like:
>>
>> /freebase/type_deprecated/
>>
>> /freebase/property_deprecated/

Easier in what way? As far as flagging something as being deprecated,
it's equally simple. In terms of retrieving data (especially in the
case of not wanting to use deprecated properties or types, I'm not
sure that "b:type":[{"id":"/freebase/property_deprecated","optional":"forbidden"}]
is any easier than
"/freebase/type_hints/deprecated":{"value":true,"optional":"forbidden"}.
But I'm certainly willing to explore other options if they'll make
life easier or more performant for people.

>> ?
>>
>    I like this idea.
>
>    Looking at the metaschema,  it's clear people have made many different
> modeling decisions at different points in Freebase?

You can say that again!

>    Would it be good to establish a preferred pattern?  As an infovore I find
> an extract of
>
> fbase:m.02834 a fbase:whatever.type .
>
>     records to be especially yummy.

In general, the preference is for using properties rather than cotypes
(especially ones with no properties) to assert things about an entity.
But as I said above, there's no reason to be bound by this if a
different pattern is better.

Jeff

> _______________________________________________
> You are receiving this message because you are subscribed to the
> Freebase-discuss mailing list.
> To post a message to the list: Freebase-discuss at freebase.com
> To unsubscribe, view archives, etc:
> http://lists.freebase.com/mailman/listinfo/freebase-discuss
>


More information about the Freebase-discuss mailing list