David Karger wrote:
>  I suppose one could argue for "simplifying" any facet by leaving out any 
> value that is tied to _every_ element in the active set---that would fix 
> this one pathological case, 
Yes, our case is a bit unusual. But the motive of my starting point is 
to allow users to have an AND relationship between the conditions. I 
guess another thing that I could try, is to associate the property Px 
with a RDF collection, containing A and B. Then they would be able to 
find resources all sharing that collection of values.

The last thing I want to happen is when I show Exhibit to my users, they 
got confused by seeing another A in facet Py.

Thank you very much for your helps. I liked your example of searching 
for the spicy food recipe. :-)

---
Jun

> and might even be useful for less 
> pathological scenarios: for example, after I narrow my recipes down to 
> "ingredient contains chili peppers" it might be pointless for my 
> "flavors" facet to still contain "spicy" as an option. 
>
> Jun Zhao wrote:
>   
>> Hi David,
>>
>> No, I am not sure about this. If I have an item I1, and has two 
>> properties Px, Py, where Py is a copy of Px but with a different name, 
>> then I should have:
>>
>> I1   Px   A
>> I1   Py   A
>> I1   Px   B
>> I1   Py   B
>>
>> When I select Px=A for I1,  then in the facet of Py, I should still see 
>> the value of B AND A, because item I1's property Py has the values of A 
>> and B. But in facet Py, I don't want to show the "A" to the users, but 
>> only the "B", to help them to continue to filter their records.
>>
>> That's my understanding about how the facet browsing works in Exhibit. 
>> Please correct me if I am wrong or I misinterpreted what you meant as 
>> making a copy of the property
>>
>> Thanks,
>>
>> Jun
>>
>> David Karger wrote:
>>   
>>     
>>> Well, if you want to carry the hack further, make a copy of the property 
>>> and give it a different name. Then create a facet on the copied 
>>> property.  Since it is a different facet it will get filtered against 
>>> the first.
>>>
>>> Jun Zhao wrote:
>>>   
>>>     
>>>       
>>>> Hi David,
>>>>
>>>> Thank you for your prompt answer.
>>>>
>>>> I have tried this. However, this gets confusing if users want to 
>>>> construct an AND in two steps. The second facet will not get filtered 
>>>> when you make a selection in the first facet, because they are copies of 
>>>> one another.
>>>>
>>>> Thanks,
>>>>
>>>> Jun
>>>>
>>>>
>>>> David Karger wrote:
>>>>   
>>>>     
>>>>       
>>>>         
>>>>> One hack that seems to work is to create two copies of the same 
>>>>> facet---then you get a conjunction of what is selected within each.  
>>>>> This isn't as elegant as AND over a single facet, but it is more 
>>>>> powerful because you can AND two separate "ORs", as in selecting color 
>>>>> "red or green" AND "blue or yellow".
>>>>>
>>>>> Jun Zhao wrote:
>>>>>   
>>>>>     
>>>>>       
>>>>>         
>>>>>           
>>>>>> Hello there,
>>>>>>
>>>>>> Now in Exhibit if I select multiple conditions in one facet, I will get 
>>>>>> everything that satisfies any of the filtering conditions. I am 
>>>>>> wondering whether there is anyway I can make these multiple conditions 
>>>>>> conjunctive, to get me everything that needs to satisfy all the 
>>>>>> filtering conditions.
>>>>>>
>>>>>> Many thanks,
>>>>>>
>>>>>> Jun
>>>>>> _______________________________________________
>>>>>> General mailing list
>>>>>> [email protected]
>>>>>> http://simile.mit.edu/mailman/listinfo/general
>>>>>>   
>>>>>>     
>>>>>>       
>>>>>>         
>>>>>>           
>>>>>>             
>>>>> _______________________________________________
>>>>> General mailing list
>>>>> [email protected]
>>>>> http://simile.mit.edu/mailman/listinfo/general
>>>>>   
>>>>>     
>>>>>       
>>>>>         
>>>>>           
>>>> _______________________________________________
>>>> General mailing list
>>>> [email protected]
>>>> http://simile.mit.edu/mailman/listinfo/general
>>>>   
>>>>     
>>>>       
>>>>         
>>> _______________________________________________
>>> General mailing list
>>> [email protected]
>>> http://simile.mit.edu/mailman/listinfo/general
>>>   
>>>     
>>>       
>> _______________________________________________
>> General mailing list
>> [email protected]
>> http://simile.mit.edu/mailman/listinfo/general
>>   
>>     
> _______________________________________________
> General mailing list
> [email protected]
> http://simile.mit.edu/mailman/listinfo/general
>   

_______________________________________________
General mailing list
[email protected]
http://simile.mit.edu/mailman/listinfo/general

Reply via email to