D3 is extremely powerful, and therefore complicated as heck.
These two qualities tend to be coupled.
Good point. By the way, Atlas uses D3, because I rewrote the earlier graphs that were using flot for the exact same reason.
The advantage of Shiny, for example, is that we can keep using our R code and expertise and the powerful ggplot2 graphing library.
I do quite like R.
And maybe there are other great frameworks out there that we haven't looked at yet.
From my exposure, I've felt that it was between flot and D3, with D3 increasing in popularity lately. It seems that most of my friends in UC Berkeley/Bay Area startups use one of these two. And R is predominantly used by academics.
But I agree that there may be another framework that we haven't found. I'm personally would like D3 > R+ggplot2, but I use R more (so I guess that says something too, doesn't it...).
The good thing is that we don't have to decide on a framework today, we should just talk about reasons for or against picking any of them, because none of them seems to work without JavaScript.
I really think that we should incorporate JS. The additional functionalities we could add would really provide a lot of value to many people. (We can provide a minimal functionality for people who do not have JS enabled.)
And we really shouldn't write our own.
+1
Cheers, Linda