[Rd] R graphics device for flash apps

[Byron Ellis]

Hi Jeffrey,

On 7/6/07, Jeffrey Horner <jeff.horner at vanderbilt.edu> wrote:
> Byron,
>
> I just read your blog (statcomp.blogspot.com, linked from your other
> post) and rand across your idea of an R/Flash graphics device. I've also
> been giving this some thought because of this amazing interactive flash app:
>
> http://tools.google.com/gapminder
>
> and others from http://www.gapminder.org/.

Ah, yes. Gotta love Hans Rosling (the guy can swallow swords too!).
:-) When I proposed the idea I was thinking more of the Adobe Flex
sort of applications where you're pushing more static graphics as SWF
through to the Flash side of things. If you poke around Google Video
for the Developer's Day stuff, one of the Flex guys had some nice
graphs in a Flex application. He was showing off Gears and it was a
pretty trivial CRM application, but you can imagine R coming into play
pretty easily.

>
> I would love more than anything to work on this. The only hitch is that
> I'm so booked up with work (not to mention a new RApache release soon)
> that I don't have the time. And believe me, getting even minimal

Yeah, I know the feeling. The RExecServer stuff I've been doing has
been an extra-curricular activity just to do some different
programming for a bit. Probably when I finally make the commits I've
been threatening for a month now into the Bioconductor repository. :-)

> functionality to work, like 2D images with pop-ups labels and such...
> would be a tremendous development effort. Not to mention trying to keep
> up with an implementation for different OS's and flash player versions.

I think the best way to start would be implementing the traditional
graphics device to push SWF. My thought that a lot of the gadgetry
would be done by the UI designer folks and that the R generated stuff
would be piped in by a surrounding Flash application. After that would
come adding non-traditional elements to the plots, likely via another
interface. I think Ming would probably be able to handle it, but
compilation is non-trivial making portability difficult.

>
> So I have questions for the R development community:
>
> 1. Are there others interested in this? Is anyone working on an
> implementation? If not, is it too niche to even consider doing?
>
> 2. Could a project like this obtain funding, any kind, not just academic?

Good question. My guess is that it would be fundamentally tied to #3.
(I wouldn't have high hopes for academic funding though...) But, hey,
who knows? Only one way to find out I suppose.

>
> 3. How does the community feel about a closed-source version?
> There are other closed-source packages available for R (the only one I
> know of is xlsReadWrite), so it seems to be a reasonable question.
>
> Comments, criticisms, flames... are all welcome. Email me off-list if
> you'd like.
>
> Jeff
> --
> http://biostat.mc.vanderbilt.edu/JeffreyHorner
>


-- 
Byron Ellis (byron.ellis at gmail.com)
"Oook" -- The Librarian