David Huynh wrote:
> Lovely!
>
> I think in this case clicking on a facet choice should not filter out 
> points but rather highlight the points that match the facet selection. 
> It's a bit disorienting after clicking on Word Series Winner "yes"--I 
> didn't expect the plot to change shape.
>   
Note there are two different kinds of change: changing the point set, 
and changing the axes.  Both happen when I pick a set of points that 
spans less years.  They need to be considered separately.  Switching the 
axes seems to be a bit like rezooming a google map when the set of 
points changes.  Probably rather disruptive of context if, e.g., i've 
got my finger on a spot on the screen. 

I like davidh's idea of highlighting points but I think that erasing 
points has its place too.  perhaps the effect should be choosable using 
another attribute in the facet.
> I'm also getting lots (thousands) of errors in Firebug:
>     plot._dataSource has no properties, timeplot_bundle.js (line 533)
>     var v = plot._dataSource.getValue(t);
> But I think you've mentioned these before.
>
> David
>
> Ryan Lee wrote:
>   
>> I've added a Timelot extension for Exhibit, much like the Timeline and 
>> mapping extensions currently in use.  It's in a developmental stage.
>>
>>    http://ryanlee.org/2008/03/rsy/rivalry.html
>>
>> We're aware of some of the logged errors and their source.  The main 
>> issue we'd like to resolve now before moving forward is how best to 
>> model time series information in RDF.  So far we've come up with a 
>> couple scenarios.
>>
>> The first case is illustrated in the above example, where every point on 
>> the plot is also an item in the Exhibit database.  As you can see, this 
>> allows you to modify the graph and move around decades just by making 
>> facet selections.
>>
>> The second case is one where modifying series via facets is of little 
>> use.  Far better for scaling (keeping the total number of items down) 
>> would be to encapsulate the entire time series in some sort of data 
>> structure, like a comma separated list of tuples in a string, attached 
>> to an item.
>>
>> I suspect orienting a model around time series makes that particular 
>> grounding difficult to use in other situations.  Maybe that's just the 
>> price.  Thoughts?  Other models?
>>
>> Interested parties can watch the wiki page below as things stabilize.
>>
>>    http://simile.mit.edu/wiki/Exhibit/2.0/Timeplot_View
>>
>>   
>>     
>
> _______________________________________________
> 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