May 2, 2012, 7:10:18 PM5/2/12
Thanks for the question. There are a lot of reasons why we switched
from a graphical user interface (Cube 0.1) to an application
programming interface (Cubism 1.0). The big one is that an API has a
much higher power-to-weight ratio: we can enable more expressive
dashboards more easily with a simple library than we can trying to
develop a comprehensive GUI.
Cubism.js provides a small set of orthogonal components (metrics,
charts, stylesheets, etc.) that are assembled via code. Cubism's
charts are much improved over those previously built-in to Cube, and
they are also much easier to customize and extend. You can define new
chart types, new stylesheets, new forms of interaction, combine cubism
charts with other HTML, etc.
You can still implement a graphical constructor on top of Cube +
Cubism, if you like. Or you could use a wiki to let people edit static
files that embed Cubism charts. Perhaps we'll release an example of
this in the future.