Brendan Taylor wrote:
> On Tue, Jan 02, 2007 at 11:51:11PM -0500, David Huynh wrote:
>
>> It's something on my to-do list--to allow multiple hops in facets. It's
>> a little trickier to implement than single hops.
>>
> So multiple hop expressions are only supported in lenses?
>
That was true 5 minutes ago. But now you can have multiple hops if you
link against this
http://simile.mit.edu/repository/exhibit/trunk/src/webapp/api/exhibit-api.js
and let me know if there is any bug :) I'll move it over to the release
site if you can't find anything wrong.
>>> 2. If I just use ex:facets=".sections", I can select Combinations
>>> containing a certain Section. If I select "s1" and "s2", it will show me
>>> combinations containing s1 OR s2. I want it to show me combinations
>>> containing s1 AND s2. (ex:facets=".sections, .sections" just confuses
>>> Exhibit)
>>>
>> You're right. The current behavior is OR. I wonder how the facet's UI
>> must be changed to indicate that the behavior is AND. Any suggestion?
>>
> Hadn't really thought about it, I was hoping you had. :)
>
I had... but nothing good came :)
> Separate facets already AND together, maybe presenting the same field
> several times (like ex:facets=".sections, .sections") would make the most
> sense. That would take up a lot of space and limit the number of things
> a user could AND on though.
>
> First things that come to mind:
>
> - "all of" or "any of" beside the name of the facet
> - "<" or ">" to indicate whether selecting an item will decrease or increase
> the number of results
> - an intersection or union symbol somewhere (ok, maybe not)
>
Sometimes the semantics is self-evident in the facet's name... but
explanations for this kind of thing tend to get really wordy.
>> I'm glad you're asking these questions because it's signaling me that
>> those features are indeed needed. Is Exhibit still useful to you for the
>> time being without those features?
>>
> Now that I've got a better idea about what I can do with Exhibit I'll
> play around with it to see what works best for what I'm trying to do.
>
> I think I can simulate multiple hop facets if I denormalise the JSON. (eg.
> add 'courses: [ "E E 238", ... ]' to each Set.) I was already going to do
> that for facetting on things like the length of a value array.
>
OK, no need for denormalising anymore...
> I'm pretty sure I need some facets to be intersectioned rather than
> unioned, though. Faking that would be a big mess.
>
I'll continue thinking...
David
_______________________________________________
General mailing list
[email protected]
http://simile.mit.edu/mailman/listinfo/general
