[R-gui] Two distinct meanings of GUI

Bob Cain arcane at arcanemethods.com
Thu Nov 18 07:48:02 CET 2004


There seem to be two distinct meanings being given to the 
term GUI in realationship to R as used in this forum.

Lately people have been talking about the GUI that surrounds 
the use of the interpreter to write programs that solve 
problems.  The second, which has been more what's been 
talked about in the six months or so I've been here is the 
API that R coders can use to _build_ GUI's for applications 
implemented in R.

I think good arguments can be made for keeping things fairly 
simple in the first case but it is the second case, which I 
think is of far more importance, that is not going to be 
simple no matter what.  Considerations other than simplicity 
dominate that discussion.

Despite its origins, if you only consider the R language, 
sans libraries, it is one of the finest interpreted 
languages going and yields nothing general to its primary 
application.  I would like to see the use of R spread in 
other directions.  I'd like to see a digital signal 
processing library developed with fast, compiled primatives 
(FFT, root solvers, etc.) to displace Matlab.  (I only wish 
S/R designers had been wiser in their treatment of array 
index origin then Matlab but it appears far too late to fix 
that oversight.)

There are also audio applications I would like to write that 
employ this (yet to be ported) DSP functionality and which 
present to the end user nice, usable button, slider, 
selector, waverform, etc., interfaces such as he is used to 
with most modern applications.  Coming from a background 
that includes APL and various compiled languages I simply 
love the design of the R language and think it is time for R 
to grow into new arenas beyond statistics.

It seems to me that some of the people in recent discussions 
are discussing one use of GUI while others are discussing 
the other use and I'm just hoping to clarify the issues so 
that it is clear what the arguments presented apply to.


Bob
-- 

"Things should be described as simply as possible, but no 
simpler."

                                              A. Einstein



More information about the R-SIG-GUI mailing list