[Rd] graphics library fltk

Duncan Murdoch murdoch@stats.uwo.ca
Tue, 01 May 2001 13:49:31 -0400

On Tue, 1 May 2001 15:50:28 +0100 (GMT Daylight Time), you wrote in
message <Pine.WNT.4.31.0105011543000.1176-100000@gannet>:

>On Tue, 1 May 2001, Erich Neuwirth wrote:
>> we still do not have dynamic graphics in R natively.
>We do not intend to: that's what things like Duncan Murdoch's Rgl and
>Rggobi are for.  It's called not re-inventing the wheel.

I guess there's a question of what "natively" means.  I don't think
dynamic graphics should be in the base, but I do think it should be
possible to install a package to get access to them.  That's what I'm
aiming for with RGL.

It's considerably more difficult for the user to install a separate
system (Java, GGobi, or whatever).  Just about all free software
packages are fairly unfriendly about installation, so learning two
sets of rules (both R's and the other's) is a substantial barrier to
entry.  There are also the different conventions about user
interaction with the 2nd package to learn.

To help people writing such packages, it would be nice if the
underpinnings to allow dynamic graphics were in the base, e.g. exposed
data in graphics objects that can be changed, and the ability to
request a redraw of the current plot once the data have changed.
("Data" here is not just the points being plotted, I'm thinking of all
the parts of the graph:  the axes, labels, transformations, etc.)

We don't want to re-invent wheels, but I think we do want to
facilitate experimentation with graphics. 

>The hard bit is to find a way to describe how to use this from
>an R interface.  Things like the placement of widgets are in my experience
>much easier with visual tools.

That's definitely true, but I don't see why a package that provided
some visual tools couldn't be written.  I'm not asking someone to
write it right away (nor volunteering!), but I think it's something
that would be easier if the graphics model changed a bit, and I know
that's something that Ross and Paul are thinking about a lot.

Duncan Murdoch
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch