[Data-modeling] Tweaks to the food domain

[Ed Laurent]

It looks like some of the Beverage properties are duplicating what should be
associated with Consumer_product and Alcoholic_beverage.

I would describe Beverage with very generic properties. For example, "Jack
Daniel's Whiskey Single Barrel 94@ 750ML" would be a Consumer_product linked
to "Jack Daniel's Whiskey Single Barrel 94" that is linked to Whiskey, which
is co-typed with Beverage and Alcoholic_beverage.

Similarly, "Milk" would be typed with Beverage. I could see "Skim milk" as a
topic of a "Milk type" enumerated type or go the route of adding properties
to a Milk_product type that could be co-typed within Cheese and used to
describe "Mozzarella cheese".

Another route would be to develop the schema so that the consumer product
"Fresh Mozzarella" would be typed with Consumer_product and include a link
to a Cheese CVT that includes properties such as Cheese_type, Milk_source
(e.g., cow, goat), etc.

Somehow I went from Beverage and Whiskey to Cheese...

-Ed

On Fri, Jun 6, 2008 at 7:40 PM, Jeff Prucher <jeff at metaweb.com> wrote:

> There are bunch of types for various beverages in the food domain, some of
> which could use a little fixing up. (Most of these were also made by me --
> mea culpa!)
>
> 1) Beverage <http://www.freebase.com/tools/schema/food/beverage>. I'm
> wondering if this type is even necessary. The instances are a mix of
> general
> beverage varieties ("Cocktail") and specific varieties ("Newcastle Brown
> Ale"), so at the very least this should be normalized and cleaned up. And
> both them "manufacturer" and "place of origin" properties seem redundant
> with some of the more fully-fleshed out beverage schemas, like beer and
> wine.  And the "color" property is probably pointless.
>
>
> 2) Whisky: This has a similar problem as beverage, in that the instances
> include both distilleries and products. I think that a simple fix would be
> to add a "distillery" property (and type), and reserve the "whisky" type
> for
> the specific whiskys (similar to the beer/brewery relationship).  There's
> probably a lot more that could be done with whiskies, but for now I'd just
> like to clean up the current model.
>
> What do you all think?
>
> Jeff Prucher
>
>
> _______________________________________________
> 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/20080608/d3eba234/attachment.htm