It's been suggested I use R, I suppose because my employer is too skint to
pay
for a copy of Splus.
But.... it doesn't seem to be stable. I was trying to do some ordinary
regression
and stepping through a model. I was trying to set up a model
('kitchen_sink') in
which to put all my model parameters. Then I wanted to use step() to see what
it
would reduce the model to. I couldn't even get started though. I would
specify
the same model again and again, with just a few of the possible variables, and
sometimes lm() would fall over and return an error message, sometimes the
regression
ran fine. I got fed up and ftp'd the original tables up to a machine
running
Splus, where the models ran fine, every time, with all kinds of specifications.
Is this a normal experience? Is R buggy or what? I note the other email that
R can't handle files named x. If I can use Splus, should I just do that?
(Alternative is graphics-interface-only SPSS, and given that I've never used
it,
I'm working on remote machines, from home, with no manuals handy, that is
something I
should really wish to avoid :-).
-julii
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
If you really want help it is generally best to include real examples and to be explicit about what is not working. If on the other hand you just want some reassurance that Splus is indeed a commerical package it is. If your question is whether R is reliable, I believe that there are a large number of users that are quite happy with its performance. There are also some that aren't. We do come with a money back guarantee though ;-) On Tue, Dec 05, 2000 at 03:41:16PM +0000, J.Brainard at uea.ac.uk wrote:> > It's been suggested I use R, I suppose because my employer is too skint to pay > for a copy of Splus. > > But.... it doesn't seem to be stable. I was trying to do some ordinary regression > and stepping through a model. I was trying to set up a model ('kitchen_sink') in > which to put all my model parameters. Then I wanted to use step() to see what it > would reduce the model to. I couldn't even get started though. I would specify > the same model again and again, with just a few of the possible variables, and > sometimes lm() would fall over and return an error message, sometimes the regression > ran fine. I got fed up and ftp'd the original tables up to a machine running > Splus, where the models ran fine, every time, with all kinds of specifications. > > Is this a normal experience? Is R buggy or what? I note the other email that > R can't handle files named x. If I can use Splus, should I just do that?This is simply untrue. Perhaps you could read those discussions somewhat more carefully. Both R and Splus have some problems with scoping at times. This is being fixed in R, you would have to ask the folks in Seattle if they are fixing their versions. However, since they have been around longer most people know how to program around their problems. R is somewhat newer but that knowledge combined with accurate bug reports (are you indeed attempting to report a bug?) will help.> > (Alternative is graphics-interface-only SPSS, and given that I've never used it, > I'm working on remote machines, from home, with no manuals handy, that is something I > should really wish to avoid :-). > > -julii > -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- > r-help 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-help-request at stat.math.ethz.ch > _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._-- +---------------------------------------------------------------------------+ | Robert Gentleman phone : (617) 632-5250 | | Associate Professor fax: (617) 632-2444 | | Department of Biostatistics office: not yet | | Harvard School of Public Health email: rgentlem at jimmy.dfci.harvard.edu | +---------------------------------------------------------------------------+ -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help 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-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
J.Brainard at uea.ac.uk writes:> It's been suggested I use R, I suppose because my employer is too > skint to pay for a copy of Splus. > > But.... it doesn't seem to be stable. I was trying to do some > ordinary regression and stepping through a model. I was trying to > set up a model ('kitchen_sink') in which to put all my model > parameters. Then I wanted to use step() to see what it would reduce > the model to. I couldn't even get started though. I would specify > the same model again and again, with just a few of the possible > variables, and sometimes lm() would fall over and return an error > message, sometimes the regression ran fine. I got fed up and ftp'd > the original tables up to a machine running Splus, where the models > ran fine, every time, with all kinds of specifications. > > Is this a normal experience? Is R buggy or what? I note the other > email that R can't handle files named x. If I can use Splus, should > I just do that?We haven't heard of problems of that magnitude lately. In general, we try to fix problems as soon as they can be identified, but you're not giving us much to go on with that kind of report. - which version of R? - which operating system? - which commands are you trying to run? - what are the error messages and other symptoms of misbehavior? -- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk) FAX: (+45) 35327907 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help 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-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
J.Brainard at uea.ac.uk wrote:> It's been suggested I use R, I suppose because my employer is too skint to pay > for a copy of Splus. > > But.... it doesn't seem to be stable. I was trying to do some ordinary regression > and stepping through a model. I was trying to set up a model ('kitchen_sink') in > which to put all my model parameters. Then I wanted to use step() to see what it > would reduce the model to. I couldn't even get started though. I would specify > the same model again and again, with just a few of the possible variables, and > sometimes lm() would fall over and return an error message, sometimes the regression > ran fine. I got fed up and ftp'd the original tables up to a machine running > Splus, where the models ran fine, every time, with all kinds of specifications. > > Is this a normal experience? Is R buggy or what? I note the other email that > R can't handle files named x. If I can use Splus, should I just do that? > > (Alternative is graphics-interface-only SPSS, and given that I've never used it, > I'm working on remote machines, from home, with no manuals handy, that is something I > should really wish to avoid :-). > > -julii > -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- > r-help 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-help-request at stat.math.ethz.ch > _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._Hi Julii I've been working with R since version 0.49 and I'm very happy with it ! I started using R before using S+ in my Linux box. I had a S+ license during 1998 and I'm now looking again at S+ because I want to use the S+SpatialStats. I think there's some differences but most of what I do I do it in R (I'm a Marine Biologist) and for "commom" statistic analysis you have everything you need in R. If you want to do something more specific you can allways write your own functions and make them available to the R community :-) The main differences for me are the manuals, but you can allways use the S+ manuals with some changes, the trellis graphics and some modules (SpatialStats for example). Anyway as someone stated you allways get a money back guarantee and the differences are 3000 EUR for a Linux License ... Best regards EJ -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help 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-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On 05-Dec-2000 J.Brainard at uea.ac.uk wrote:> It's been suggested I use R, I suppose because my employer is too skint to > pay for a copy of Splus. > > But.... it doesn't seem to be stable.[snip] I've used both and I prefer R for a couple of reasons: 1) Support: This list is great. It's never failed to come up with an answer to my problems almost instantly. And it's usually the core developers that answer. I had a strange problem with alignment of text in Splus and it took a week or 10 days for the people at MathSoft to come up with an answer. 2) Swiftness: SPlus is heavy and lumbering, while R is nimble. R runs fine on my old Thinkpad 701 with 300 MB hard drive and a 75 MHz 486 under Linux. Want to try and install SPlus on that machine? ______________________________________________________________________ Stuart Luppescu -=-=- University of Chicago $(B:MJ8$HCRF`H~$NIc(B -=-=- s-luppescu at uchicago.edu http://www.consortium-chicago.org/people/sl/sl.html PGP Public Key: www.consortium-chicago.org/people/sl/pubkey.asc ICQ #21172047 AIM: psycho7070 Money is better than poverty, if only for financial reasons.>> Sent on 05-Dec-2000 at 17:17:52 with xfmail-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help 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-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
J.Brainard at uea.ac.uk Writes:>It's been suggested I use R, I suppose because my employer is too skint to pay >for a copy of Splus.Or, maybe, he just supports open-source software ... some of us do. I can't see how an 'ac.uk' employer would pay more than a few quid for a copy of Splus through an academic licensing deal.>But.... it doesn't seem to be stable. I was trying to do some ordinary >regression [stuff deleted]I am sure that other on the list will ask you to give more details about your machine, OS, R versions, dataset, commands that caused the problem, error messages, &c.>Is this a normal experience?Not mine. I find it stable.> Is R buggy or what?Getting less so. I am very surprised that lm() is giving you problems.> I note the other email that >R can't handle files named x.One of us is confused. Where did that come from?> If I can use Splus, should I just do that?If you feel happier with it. R is lighter and faster than Splus. R support (through the lists) is excellent.>(Alternative is graphics-interface-only SPSS, and given that I've never used it, >I'm working on remote machines, from home, with no manuals handy, that is >something I >should really wish to avoid :-).SPSS is a 'package' rather than an environment/language like R or S. It is pretty good (stable, fast, easy to use) but limited in the way 'packages' usually are. I don't think that there are any "graphic- interface-only" versions of SPSS. The GUI, in SPSS, is a syntax builder just like the GUI in Splus. The command language is a bit idiosyncratic but you can use the GUI to help you see how commands are structured. SPSS has good on-line help and the manuals are good (you can get them in most academic bookstores). SPSS is much simpler than R/S for data- management jobs. SPSS is a venerable package and will do the job for most users most of the time. It is very cheap under academic license. Of course, there is some pride in being an SPlus, and by implication R, user as these systems have the reputation of being tools for illuminati. Just my tuppence. Mark -- Mark Myatt -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help 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-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
J.Brainard at uea.ac.uk writes:> It's been suggested I use R, I suppose because my employer is too > skint to pay for a copy of Splus.Hi Julii, you're not still at UEA are you? Haven't heard of you for a while. Anyway, there's a small group of us using R in the Centre for Statistics and I've just updated our installation. I've also asked for the HPC to do the same, but in the mean time, I think I could give you access to our systems if necessary. [For list reference ;-), I believe UEA HPC is using 0.9x, while the Macs were running 0.6x until recently because that was the latest available.] -- MJR (Not an official statement) Please note changes of address http://stats.mth.uea.ac.uk/ -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help 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-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._