Re: Ontology for Restaurants

Peter, your observations are of course valid, but consider that we 
assume an "open world", so that some knowledge may be left unspecified 
(of course one cannot model the whole world in which we live), and one 
can decide a level of "granularity" of the information provided. In 
addition, I can't understand if the approach you are working on aims to 
solve the issues you mentioned.

CL

On 13/07/2013 22:03, Peter Brooks wrote:
> Given that almost every living thing and many thousands of derived 
> compounds as well as inorganic compounds can be ingredients, an 
> ontology of ingredients would be quite difficult to put together. Not 
> only that, but many ingredients of recipes are recipes themselves - in 
> a moussaka, for example, bechamel sauce is a dish all on its own, as 
> well as an ingredient. Then Solanum esculentum is known as aubergine, 
> egg-plant, brinjal, melongene, or guinea squash and, is a species of 
> nightshade - so a warning that 10% of people have an allergy to it 
> might be useful. A good ontology would also note that it's a good 
> source of potassium. It might also be wise to have a warning that this 
> particular ingredient contains the largest concentration of nicotine 
> than any other edible plant. Quite a few different properties, and 
> that's before you consider whether the mode of preparation is a 
> relationship or an ingredient. I wonder how an ingredient ontology 
> would deal with a copper bowl that is, apparently, essential if you 
> want the best whipped egg-white.

Received on Saturday, 13 July 2013 20:56:54 UTC