You are correct, that with many of these apps, the main bottleneck is 
getting the browser to update the dom (whether or not a piece of it is 
visible).  There is indeed an approach to dealing with this problem, 
essentially to keep track of which elements of the view are actually 
visible at a given time, and draw only those images.   A good example of 
this approach can be found in the following paper:
http://doi.acm.org/10.1145/1294211.1294235
(I haven't found an openly-accessible copy).  I saw the system demoed at 
the conference and it was impressively snappy, even on a very large data 
set. 

Of course, this isn't a minor feature; it requires significant work to 
build a system this way.  Hopefully someday it can be done for timeline 
(and exhibit)

-David

Ron Van den Branden wrote:
> Hi,
>
> We've been experimenting with Timeline as an interface for an electronic 
> text edition of drafts and notebooks used by literary authors. In 
> principle, we think it could be the basis for a very promising 
> visualisation of search results on a time-based document collection.
>
> However, a demo application with a data file containing over 3200 events 
> proved fairly slow already (probably too slow for researchers to 
> comfortably work with the timeline). Rather than its size (over 3 Mb), 
> the Javascript-driven drawing of the sheer amount of events seems to be 
> the bottleneck (deleting all events' text didn't improve performance). 
> Since these sample data consist of only one fourth of the intended 
> amount of data of the total edition, my question is: what would be the 
> most sensible way to have Timeline deal with huge data collections? Is 
> there a way to 'cut off' drawing of events to just the current visible 
> 'time window'?
>
> Kind regards,
>
> Ron
>
>
>
> _______________________________________________
> 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