Thanks to Prof Ripley, Arne, Andy & Bill for unambiguous suggestions! Linux box is on order. I'll take notes on our experience & post a follow-up in a few weeks. May be useful to other folks stuck in the Windows world.
-Jim
At 06:24 AM 1/29/2004, Pikounis, Bill wrote:
Jim, I would really like to reiterate Professor Ripley's and Arne Henningsen comments. The problem goes for any analytic software or system you might want to use, not just R. My impression is that at least for part of it, you want the individual users to use R as they would on their own desktops. (If that is not the case, much of the rest of this note is pure FYI.) Even in its most advanced 2003 Server edition, Windows is simply not designed to be a multi-user system. Sure, it can reliably host a web server that may need to run quick bursts of R batch-type jobs ("analytics") and return results to a client (e.g. web browser), but that does not sound like what you are looking for (at least in part). And beyond the technical limitations, use of Windows Terminal Server (Remote Desktop) / Citrix, etc. will cost much money and implementation hassle and probably even legal headaches. We have had colleagues here at Merck (over my and Andy Liaw's disbelief) that have tried to shoehorn Windows this way, and even the speed of single, small jobs by 1 logged-on took longer on the server than on their much less powerful laptops.
A Linux solution is very flexible, in our experiences (we have Windows XP as corporate desktop standard). As stated, with Samba, you can map directories that look like just another drive in Windows Explorer. Printing is just as transparent in either direction. VNC (Virtual Network Computing) is very, very nice to provide the individual user's Linux environment as just another window on their Windows desktop. With the free utility of "autocutsel", clipboards can be synchronized for ease of cutting and pasting. And KDE, one of several window manager analogues to Windows, is very sophisticated and shares a lot in common with the Windows GUI from a user operations standpoint. While it may sound like a hassle to get up and running now if your shop is currently "99% Windows", the benefit will absolutely be clear later.
Hope that helps, Bill ---------------------------------------- Bill Pikounis, Ph.D.
Biometrics Research Department Merck Research Laboratories PO Box 2000, MailDrop RY33-300 126 E. Lincoln Avenue Rahway, New Jersey 07065-0900 USA
Phone: 732 594 3913 Fax: 732 594 1565
> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Arne Henningsen > Sent: Thursday, January 29, 2004 3:45 AM > To: Jim Porzak > Cc: [EMAIL PROTECTED] > Subject: Re: [R]Running R remotely in Windows Environment? > > > Hi, > > I also suggest to use a Linux Server. You can work on this > machine via ssh > (e.g. with PuTTY) and transfer the input and output files > with scp or a samba > server (which is easy to install and very convenient to use > for windows > users). > > Arne > > On Thursday 29 January 2004 08:53, Prof Brian Ripley wrote: > > On Wed, 28 Jan 2004, Jim Porzak wrote: > > > We are considering setting up a fast, RAM loaded machine > as an "R-server" > > > to handle the big problems not suitable for individual > desktops and, > > > also, to process ad hoc analysis requests via our portal. > We are 99% a > > > Windows shop, so first choice is a windows server. We'll > use (D)COM for > > > the portal interface and understand that. > > > > > > What has me stumped is how to easily interface individual > analyst's > > > Windows desktops to the R-server. I haven't seen anything in the > > > archives, but I can't imagine this hasn't been done. What > am I missing? > > > > R is not designed to be client-server on Windows. People I > know who do > > this use Windows Terminal Server or Citrix. > > > > I would question the value of this approach. Unless you > propose to run > > 64-bit Windows, a `RAM loaded' machine isn't `loaded', and > R under Windows > > handles large amounts of memory much less effectively than > under Linux. > > 64-bit Windows is uncharted territory for R, whereas 64-bit > Unix/Linux is > > well trodden. > > -- > Arne Henningsen > Department of Agricultural Economics > University of Kiel > Olshausenstr. 40 > D-24098 Kiel (Germany) > Tel: +49-431-880 4445 > Fax: +49-431-880 1397 > [EMAIL PROTECTED] > http://www.uni-kiel.de/agrarpol/ahenningsen/ > > ______________________________________________ > [EMAIL PROTECTED] mailing list > https://www.stat.math.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide! > http://www.R-project.org/posting-guide.html > >
------------------------------------------------------------------------------ Notice: This e-mail message, together with any attachments,...{{dropped}}
______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html