From s-sender@utstat.toronto.edu Mon Jun 17 11:56 PDT 1996 Received: from utstat.toronto.edu (utstat.toronto.edu [128.100.73.1]) by cs.sfu.ca (8.7.5/8.6.12) with SMTP id LAA05615 for ; Mon, 17 Jun 1996 11:56:09 -0700 (PDT) Received: by utstat; Mon Jun 17 12:59 EDT 1996 Received: from dicsmss1.jrc.it (dicsmss1.jrc.it [139.191.1.65]) by hermes.stat.wisc.edu (8.6.12/8.6.12) with SMTP id LAA14793 for ; Mon, 17 Jun 1996 11:59:37 -0500 Received: from ei.jrc.it (gaia.ei.jrc.it) by dicsmss1.jrc.it (4.1/EB-950131-C) id AA08625; Mon, 17 Jun 96 19:03:37 +0200 Received: from rem1 by ei.jrc.it (4.1/EI-3.1m) id AA04630; Mon, 17 Jun 96 18:58:14 +0200 Posted-Date: Mon, 17 Jun 1996 18:58:12 +0200 Sender: gregoire.dubois@jrc.it Message-Id: <31C58EA4.102F@ei.jrc.it> Date: Mon, 17 Jun 1996 18:58:12 +0200 From: GREGOIRE DUBOIS Organization: ENV. INST./ TP 321 /JRC /Ispra/ Italy X-Mailer: Mozilla 2.0 (X11; I; AIX 2) Mime-Version: 1.0 To: s-news@stat.wisc.edu Subject: SPATIAL STATS & GIS Content-Transfer-Encoding: 7bit X-Lines: 27 Status: RO Content-Type: text/plain; charset="us-ascii" Content-Length: 934 Greetings, I thought the following link could be of your interest. http://java.ei.jrc.it/rem/gregoire/ is the homepage for the GIS & GEOSTATISTICS mailing list. Called ai-geostats, this mailing list serves as a forum where GIS (Geographic Information Systems) users and statisticians can exchange information about spatial data analysis and geostatistics. More than 800 subscribers, coming from around 50 different countries, are sharing experience and informations in very different fields but all related to spatial data analysis. Because Splus has made a big step by developing Splus spatial stats, I thought this was the right place to post this announcement. Best regards to all Gregoire -- Gregoire Dubois (PhD student) Tel. 39-332-78.99.44 Joint Research Centre Fax. 39-332-78.54.66 Environment Inst. TP 321 Email: gregoire.dubois@ei.jrc.it I-21020 Ispra (Va), ITALY URL: http://java.ei.jrc.it/rem/gregoire/ From s-sender@utstat.toronto.edu Mon Jun 17 12:19 PDT 1996 Received: from utstat.toronto.edu (utstat.toronto.edu [128.100.73.1]) by cs.sfu.ca (8.7.5/8.6.12) with SMTP id MAA06774 for ; Mon, 17 Jun 1996 12:19:02 -0700 (PDT) Received: by utstat; Mon Jun 17 13:21 EDT 1996 Received: by cortex.vis.caltech.edu (8.6.12/VIS:4.48) id KAA04347; Mon, 17 Jun 1996 10:21:46 -0700 From: larry@vis.caltech.edu (Lawrence Hugh Snyder) Date: Mon, 17 Jun 1996 10:21:46 -0700 Message-Id: <199606171721.KAA04347@cortex.vis.caltech.edu> To: s-news@utstat.toronto.edu Subject: X11/unix graphics Content-Type: text Content-Length: 1752 X-Lines: 43 Status: RO (S-PLUS Version 3.2 Release 1 for Sun SPARC, SunOS 5.x : 1993) (Running under openwindows/X11, using primarily motif() and postscript()) Dear Gurus/Afficionados Three trivial questions, the sort that don't matter much but keep me up at night wondering: 1) Is there a way to set par() defaults in a device-independent manner? I'd like to be able to set par(las=1). If I put that into .First, I get a 'Device not active' complaint. I can write my own versions of each graphics device that I'm liable to use (e.g., motif() <- function() { par(las=1) ; motif() } postscript() <- function() { par(las=1) ; postscript() } But that seems like a heavy-handed approach. What I'd like is something analogous to ps.options() for par() parameters. 2) I call ps.options(extra.setup="0 setlinecap") to make the postscript generated by printgraph() agree with what the motif() window shows. Seems odd to have the default postscript and motif behaviors be different ('0 setlinecap' for motif(), '1 setlinecap' for postscript). Am I missing something here? 3) I'd like to always attach a local directory called .Save whenever I start up Splus, if that directory exists. In a global first file (called via the S_FIRST environment variable) I have: if (unix("ls -d .Save > /dev/null", output=F) == 0) attach(".Save", pos=2) That's almost perfect, but if .Save doesn't exist, I still get a ".Save: No such file or directory" message from unix. Any way to suppress that message? Thanks! Please respond directly to me (to keep list traffic to a minimum.) If I get more than 2 requests for a copy of the responses, I'll summarize to the list. larry@vis.caltech.edu From s-sender@utstat.toronto.edu Tue Jun 18 15:05 PDT 1996 Received: from utstat.toronto.edu (utstat.toronto.edu [128.100.73.1]) by cs.sfu.ca (8.7.5/8.6.12) with SMTP id PAA25629 for ; Tue, 18 Jun 1996 15:04:53 -0700 (PDT) Received: by utstat; Tue Jun 18 15:39 EDT 1996 Received: by cortex.vis.caltech.edu (8.6.12/VIS:4.48) id MAA02838; Tue, 18 Jun 1996 12:39:45 -0700 From: larry@vis.caltech.edu (Lawrence Hugh Snyder) Date: Tue, 18 Jun 1996 12:39:45 -0700 Message-Id: <199606181939.MAA02838@cortex.vis.caltech.edu> To: s-news@utstat.toronto.edu Content-Type: text Content-Length: 1939 Status: RO X-Lines: 49 To: s-news Subject: X11/unix graphics--answers Thanks to all who answered. Sorry not to give credit by name; you know who you are. larry ############################################################################ Question 1) "Device-independent way to set par() defaults?": Create your own local version of the function, Device.Default(), and add the line par(las=1) to it. This function gets called anytime a device is started, and therefore your addition is applied to all devices. Gory details: * Put the current definition of Device.Default() into an ascii file: > sink("Device.Default") > Device.Default > sink() or # sink() appears broken on my 3.2 Splus, so I do this instead: > echo 'Device.Default' > Device.Default > Splus BATCH in out * Now edit the ascii file 'Device.Default': > add the line par(las=1) near the bottom of the file (or add whatever other global device defaults you'd like) > add "Device.Default <-" to the top of the file (to assign the function) * Return to Splus and execute > Device.Default <- source("Device.Default") (Ignore the resulting warning message.) A file called Device.Default, containing your customization, now appears in your local directory, and is called instead of the system version. ############################################################################ Question 2) "The setlinecap setting is different for motif() and postscript()": Others confirm the discrepency, and apparently my fix: 'ps.options(extra.setup("0 setlinecap"))'. is a reasonable one. ############################################################################ Question 3) "Test for directory existence before attaching?": if(is.dir(".Save")) attach(".Save") # Uses sh function 'test' (My original error was not redirecting error output; see the definition of is.dir().) ############################################################################ From s-sender@utstat.toronto.edu Wed Jun 26 09:46 PDT 1996 Received: from utstat.toronto.edu (utstat.toronto.edu [128.100.73.1]) by cs.sfu.ca (8.7.5/8.6.12) with SMTP id JAA10410 for ; Wed, 26 Jun 1996 09:46:45 -0700 (PDT) Received: by utstat; Wed Jun 26 11:03 EDT 1996 Received: from mail.teleport.com (ip-pdx10-03.teleport.com [206.163.122.67]) by desiree.teleport.com (8.7.5/8.7.3) with SMTP id IAA20381 for ; Wed, 26 Jun 1996 08:03:05 -0700 (PDT) Message-Id: <199606261503.IAA20381@desiree.teleport.com> X-Mailer: Post Road Mailer (Green Edition Ver 1.03a) From: Marianne Reith To: s-news@utstat.toronto.edu Date: Wed, 26 Jun 1996 08:03:20 PDT Reply-To: Marianne Reith Subject: Re: Contrasts Content-Type: text Content-Length: 738 X-Lines: 24 Status: RO Forwarding note from: morris@kodak.com 06/25/96 5:37pm -0400 OK, I'm trying to transition away from SAS so am trying to use S+ to do some simple modeling. All I want to do is estimate and test a series of contrasts in an AOV setting. I have constructed the contrast matrix, a 6x5 matrix of -1,1,and O's each column the contrast I want to make among the 6 treatments called "trtfac". Now what? I've tried modeling as aov(Y~C(trtfac,my.contrast)) and get the 5 df SS not each individual contrast. I've looked at the blue book, the yellow book, and Spector's intro and can't figure out how to do this. Help! Thanks for any aid. Jeff Jeffrey D. Morris Environmental Sciences Eastman Kodak Company Rochester, NY morris@kodak.com