I've put my answers in between the questions.  

a) I am unsure about the notion of MVCs and how they fit in your model.

 loadData seems to create a new  MVC?
 if iSPlot is loaded is this where the MVCs are stored?

I'm using the MVC class that is defined in the package MVCClass.  So when a 
data set is loaded, a model is created and a MVC object is created.  loadData 
creates a new model and a new MVC.  The MVC object, which contains the data 
set information, is stored in mvcEnv, an environment in a variable called 
MVCList. 

 We should have some, non-GUI way to 
 1) list MVCs

You can list all of loaded data sets (model) names using the function, 
getModelNames(). 

 2) delete a MVC

I don't have this yet, but I can put it in.  One thing to notice is that when 
the user decides to quit, all loaded data will be removed from the environment,
mvcEnv.  

 3) set the active MVC (that seems to be there, I just don't see
   how to do it)

Yeah, I kind of fudged on this by allowing the user to call the function, 
setActiveMVC, which is actually part of the GUI.  I can write a function that 
allows them to give a model name as the parameter to set the active MVC. 

 4) getActiveMVC - should this be the name? Should the name of a MVC 
  be the name of the dataset?

I'm not sure what the question is here.  Yes, I think the name of a MVC object 
should be the same as the name of the model (data set) because I am assuming 
that there is a one-to-one relationship between a MVC object and a model.  I 
can make a function called getActiveMVC so that the user can which data set is 
active.  Right now the user can get this information by calling 
getMVCEnvVar("activeMVC"), but I can write a wrapper function around it.

 5) on the man page for getMVCEnvVar, it would help a lot to have short 
  descriptions of what the different values mean? And, why do users get
  to see them? Are they needed or is this an interface more for internal
  use?

Basically I let them have access to the information that is stored in the 
environment, mvcEnv.  This allows the user to see the active MVC, the MVC list,
etc.  It would probably be better to determine which information the user will 
need from this environment and then write accessor functions for these pieces 
of information directly.

b) in the file testScript.R I tried to use the command line version.
 But got stopped by 

I'm not sure where you got stopped here.

c) createSPlot man page never talks about the the active MVC, possibly
  it should? This function does have some side effects in that MVC?

When creating a view, I am expecting the user is creating a view of the active 
MVC so I can put that in the man page.  Really I should remove the dataName 
parameter in createSPlot because I am expecting that the user can only create 
views of the active MVC.  This is a remnant from the GUI, where the user 
previously could choose any of the loaded data set to view. 

d) I don't see how createSpreadsheet works. I was guessing it would need
  no arguments and just plot from the activeMVC, but somehow it looks at
  all MVCs?  Perhaps we need a function to list available data sets?

Yep, you can refer to my answer to part c.  You are right that it should take 
no arguments since I am expecting to view only the active MVC - again this is 
from the GUI when the user could choose which data set to view.

e) for getEvents, is this dependent on the active MVC? Does it change for
 different views? If so, we might need to print something if the view is not
 interactive.

Yes, getEvents does depend on the active MVC in a way, but only in terms of 
the type of model and not in terms of which views have been created.  I'm 
assuming that for the iSPlot package the only model that the user will have 
available (at least at this time) is the data frame model.  Thus, 
the getEvents function will always return mouseOver, leftButtonClick, 
middleButtonClick, and rightButtonClick as the possible events for views.  You 
can set these events even without a view created - then when a view is created,
that event will already be set.

f) is there some way, with colorPoint that I could optionally specify the
  color and not get the color selection?

Not the way things are currently set, but I can add it to my list of many 
things to do.
