Atlas shared

I got several answers to my question about shared atlas libraries,
see below, where you also find Dirk Eddelbuettel's answer. I tried his
'lazy' alternative by using 'alien' on my Debian system to get  .rpm
versions of the  .deb  libraries. This worked fine. I had to rebuild
R, though.

Dirk's 'hard work' alternative was also (in essence) suggested by
Douglas Bates, Reid Huntsinger, and Camm Maguire. I haven't tried
it yet, but I may when time permits.

Roger Peng suggested that it should be possible to link with
the _static_ libraries. Haven't checked, and haven't noticed any
reaction to his suggestion either. Would be nice if it worked.

Many thanks to all who answered!

G?ran

In a FAQ on an ATLAS home page I see that shared libraries are not
supported, and indeed, when I build from source I only get static
libraries. R  needs shared libraries (right?!), so how do I get that?
With Debian (and Windows?) there are no problems, since there are
prebuilt shared libraries, but how about RedHat? (These are my three
systems.)

How do I get shared atlas libraries that work with  R  on RedHat?

You could try to fudge it and simply place the Debian libraries onto
a RH system, preferably by converting the .deb into a .rpm via the alien
program.

Or you could study the build process of Debian's atlas package and apply
it to a .rpm.  My recollection from talking to our Atlas maintainer is that
this is not a straightforward process -- so your cost/benefit analysis might
lead you to the first solution.

Dirk

--
According to the latest figures, 43% of all signatures are totally worthless.

---
 G?ran Brostr?m                    tel: +46 90 786 5223
 Department of Statistics          fax: +46 90 786 6614
 Ume? University                   http://www.stat.umu.se/egna/gb/
 SE-90187 Ume?, Sweden             e-mail: gb at stat.umu.se


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

G?ran Brostr?m <gb at stat.umu.se> writes:

Roger Peng suggested that it should be possible to link with
the _static_ libraries. Haven't checked, and haven't noticed any
reaction to his suggestion either. Would be nice if it worked.

That works fine. As far as I remember you just need to set LDFLAGS
with the relevant -L option, e.g.

LDFLAGS=-L/usr/src/pd/ATLAS/lib/Linux_PIIISSE1_2/ ./configure
make

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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

In R-1.6.0 (Linux) I notice that 'summary' on a data.frame doesn't treat
logicals nicely. I only get

    event
 Length:148939
 Mode  :logical

Is this a bug? (I expected frequencies for TRUE and FALSE, of course.)

G?ran
---
 G?ran Brostr?m                    tel: +46 90 786 5223
 Department of Statistics          fax: +46 90 786 6614
 Ume? University                   http://www.stat.umu.se/egna/gb/
 SE-90187 Ume?, Sweden             e-mail: gb at stat.umu.se


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

G?ran Brostr?m <gb at stat.umu.se> writes:

In R-1.6.0 (Linux) I notice that 'summary' on a data.frame doesn't treat
logicals nicely. I only get

    event
 Length:148939
 Mode  :logical

Is this a bug? (I expected frequencies for TRUE and FALSE, of course.)

Not a bug, but a fairly obviously missing feature... Logicals used to
be automatically converted to factors so it wasn't noticed earlier. 

Looks like a one-line change to summary.default, so we might just be
convinced to put it in for 1.6.1, even if it's not strictly a bug.

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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

"PD" == Peter Dalgaard BSA <p.dalgaard at biostat.ku.dk>
    on 03 Oct 2002 11:20:21 +0200 writes:

PD> G?ran Brostr?m <gb at stat.umu.se> writes:
    >> In R-1.6.0 (Linux) I notice that 'summary' on a
    >> data.frame doesn't treat logicals nicely. I only get
    >> 
    >> event Length:148939 Mode :logical
    >> 
    >> Is this a bug? (I expected frequencies for TRUE and
    >> FALSE, of course.)

    PD> Not a bug, but a fairly obviously missing
    PD> feature... Logicals used to be automatically converted
    PD> to factors so it wasn't noticed earlier.

    PD> Looks like a one-line change to summary.default, so we
    PD> might just be convinced to put it in for 1.6.1, even if
    PD> it's not strictly a bug.

"yes" I said instinctivly -- but then realized that S-plus 6.1
doesn't summary()ze differently either.  

So we might be more careful...

Anyway, I'd want that after the change,

  lvar <- (1:20 %% 3 == 1)
  summary(lvar) ## and
  summary(as.factor(lvar))

*be* distinguishable.
If we do the change, (I'm "make checking" it), we should make it
distinguishibly, maybe even adding the word (line) "logical", e.g., like

summary(lvar)

Mode   FALSE    TRUE 
logical      13       7 

and {used in summary.data.frame()} :

sd <- summary(data.frame(x=1:10, y=round(rnorm(10),2), F=gl(2,5),L=(1:10)>7))
sd

x               y          F         L          
 Min.   : 1.00   Min.   :-1.120   1:5   Mode :logical  
 1st Qu.: 3.25   1st Qu.:-0.790   2:5   FALSE:7        
 Median : 5.50   Median :-0.695         TRUE :3        
 Mean   : 5.50   Mean   :-0.223                        
 3rd Qu.: 7.75   3rd Qu.: 0.390                        
 Max.   :10.00   Max.   : 1.530

str(sd)

chr [1:6, 1:4] "Min.   : 1.00  " "1st Qu.: 3.25  " "Median : 5.50  " ...
 - attr(*, "dimnames")=List of 2
  ..$ : chr [1:6] "" "" "" "" ...
  ..$ : chr [1:4] "      x" "      y" "F" "    L"
 - attr(*, "class")= chr "table"

Martin Maechler <maechler at stat.math.ethz.ch>	http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO C16	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._