See here for a more recent posting on this topic.
I really like RStudio Server (and should blog about some its cool features another time). But this post is more focused: getting RStudio Server on a Mac. I had the chance to speak to Joe Cheng, one of the RStudio engineers, at R/Finance 2011 and it seems there is little demand out there for a version of RStudio Server for Mac OS X. Still, it can be done, though I cheat a little to do it.
How? Run it in Ubuntu on a VirtualBox virtual machine. Setting up the virtual machine with R and RStudio Server is surprisingly easy (no need to burn CDs!). The only quirk is that I needed to click an icon at the bottom of the Ubuntu window to make the networking “bridged” (I don’t remember the exact term and I’m not at my computer right now to check); the effect of this was to give my Ubuntu instance its own IP address so I could access it from elsewhere (being low tech, I went to whatismyip.com to get my IP address).
Following the instructions at the RStudio site, I then had an instance of RStudio Server that I could access from anywhere (even from my iPad … though it turns out I can’t really use it there). Very cool.
Note that the biggest downside I see is that a VM ties up RAM (I have ordered some more), and perhaps quite a bit if you want R to be amply supplied. As discussed previously, there isn’t necessarily a performance hit for a Mac user running R in this way. Also, even though I installed the desktop version of Ubuntu (easier to do for me), I can run it in “headless” mode, so I hardly even notice it on the Mac side. I still have most of my data on the Mac side, but I added my new IP address to the list of IPs that can access my PostgreSQL data and have no trouble getting to that from Ubuntu R using RPostgreSQL.