Speaking for me, I find the notebook ideal for making things that I want to present or show others. I think that it's not so great for:
* playing with data by running a lot of different commands.
* writing a complete python module or method.
* editing my code down to a single, useful nugget.
Rodeo and Spyder (and RStudio) basically give you a terminal, a bunch of scratch paper, and a look up guide for objects that you've instantiated and graphs that you've drawn. If you haven't ever been tempted to open up an ipython session and a text editor, you probably don't need this. If you have, then this might be worth a whirl.
I completely agree. I feel bad self-promoting here, but there is a middle way. At Sense (https://sense.io) we've combined real editor with a notebook like console log. RStudio and Rodeo, by comparison, are basically copying traditional IDEs without a comprehensive log that includes visualizations. This makes it much harder to share results.
On Sense, your scripts are just scripts (comments are markdown formatted), but when executed in a Sense engine the REPL produces a rich console log similar to a IPython Notebook. This makes the logs great for sharing, without sacrificing the ability to refactor code into libraries, run it locally, or use code-oriented version control tools.
So you're basically offering an R studio that generates a series iPython notebook-esque plots/visuals rather than displaying them one by one in the "Plot" Rstudio window? Is that a fair characterization?
Yeah, I'm curious why in the marketing materials it's compared to Eclipse and Sublime Text, when it's in fact a dead ringer for R Studio in both form and function.
FYI the view function in the dataframe tab seems to try to open the whole dataset by default rather than do what RStudio does and just subset the first 1000 rows
Doesn't seem to work at all for me, regardless of the type of variable I initialize... Using Python3.4 and builtin pyvenv etc - could this be causing any issues?
