********************************************************************************
                               File READTML.TXT
         Downloading and installing TMLEST and its interface to Splus
********************************************************************************

TMLEST is a library of S functions for the computation of the truncated maximum
likelihood for gaussian and log-Weibull case. TMLEST requires ROBETH to be loaded 
into S-Plus. TMLEST is available at the URL:

               http://www.iumsp.ch/Unites/us/msp_us.htm (=> software)

Downloading
-----------
(a) For Splus 4.x, Splus 2000 or Splus 6.x/Windows (and more recent versions), copy the following files
    in a temporary directory:

    tmlwin.exe    (object files, self extracting zip-file;    april 2006)
    tmlwin_v6.exe (the same file but for Splus 6.x or 7.x;    april 2006)
    tmlman.exe    (documentation, self extracting text files; october 2001)

(b) For R (version >= 1.7.0 and < 2.0.0) / Windows, copy the following file in a temporary 
    directory
    tmlR_win.exe  (DLL+source, self extracting zip file;     april 2006)
    tmlman.exe    (documentation, self extracting text files; october 2001)

(c) For Splus 3.1/UNIX (and more recent versions), copy the following files
    in a temporary directory (noted <your.path> below) :

    tmlsrc.tar.Z  (program sources, compressed; april 2006).

Installation for Splus or R under Windows:
------------------------------------------


1. Execute the self-extracting file 

   tmlwin.exe (or tmlwin_v6.exe for Splus 6.x/7.x or tmlR_win.exe for R)
 
   The name of the directory where the file will be "unzipped" must 
   be entered. For a standard installation, it is:

             c:\Program Files\sp2000\library
   or
             c:\Program Files\Insightful\splus6x\library
   or   
             c:\Program Files\R\rwNNN\library 

  
   Same procedure for the file tmlman.exe.

   Remark: Installation for Splus 3.x/Windows is no more available.  
 
2. To invoke the library 
   a) under Splus, use the Splus commands:

   > library(robeth,T)
   > library(tmlest,T)

   b) under R, use the R commands:
   > library(robeth)
   > library(tmlest) 

Under UNIX:

The installation of TMLEST follows the general ROBETH installation procedure described
in the ROBETH file README/genral.txt. If you have write access to the Splus HOME directory,
the most convenient way to install TMLEST is the creation of a public library. Otherwise,
create your private library and skip step 1 below.

1. Change to the Splus HOME directory:

   cd <your Splus HOME dir>/library

2. Create the new directory tmlest. Type:

   mkdir tmlest

3. Decompress the tar file:

   cd tmlest 
   zcat <your.path>/tmlsrc.tar.Z  | tar xvf -

4. Create the Splus objects. If your Splus version is 4.x or lower, type:

   mkdir .Data      (as a sub-directory of tmlest)
   Splus
   > data.restore("<your.path>/tmlest.dmp")
   > q()

   If your Splus version is 5.x or higher, step 4 and 5 (compilation) are
   achieved by typing:
   
    Splus CHAPTER
    Splus make
# If not yet done, type
#   Splus
#   > source("TMLgauss.s")
#   > source("TMLweibul.s")
#   > source("sp2000wgt.s") # Only for Splus 5.x; not required for Splus 6.x
#   > q()
    Goto step 6.

5. Compile and link (ld command) tmlfor.f to a file called tmlfor. 
   For more details, follow the hints in genral.txt. See also the
   file loadobj.com. 

6. To invoke the public library, open Splus and type:

   > library(robeth,T)   
   > library(tmlest,T)
   
7. For a private library, create a file "mylib.q" for example, containing
   the following line:
   assign("lib.loc","<the path to the library>",frame=0)
   In my case, I have: assign("lib.loc", "/users/arandria", frame=0)
   Then after opening Splus, type:

   > source("mylib.q")
   > library(robeth,T)
   > library(tmlest,T)

[End of file]