Skip to content

Hmisc and Design libraries

1 message · Frank E Harrell Jr

Frank E Harrell Jr
# 
After several nice and informative discussions with Bill Venables,
Duncan Temple Lang, Tony Rossini, Martin Maechler, and Robert
Gentleman, I am pleased to announce that I am becoming an R user
and have begun porting the Hmisc and Design libraries to R. 
I have been very impressed with the maturity of R and with its 
logical design, particularly its online documentation setup and
the way packages are downloaded and installed, not to mention the
scoping rules and other fundamentals.  Another factor that has
influenced me is the excellent book "Rebel Code" by Glyn Moody
which chronicles the open source movement and Linux.  I wish
to especially thank Bill Venables for his friendly wisdom.

My guess is that porting the libraries can be largely
accomplished in the next 4 months (there are about
350 functions), especially if any experienced R programmers
who want to use the libraries would like to help.  I am
in the process of setting up an ssh-accessed CVS repository
for each library so that functions can be checked in and
out.  Anyone wanting to help will of course get plenty
of acknowledgement.  I have already debugged about 1/4
of the Hmisc library under R.

Descriptions of functions in my libraries may be found
at http://hesweb1.med.virginia.edu/biostat/s/splus.html

As I will be maintaining the libraries for S-Plus (both
Windows and Linux/Unix versions, S version 3 and 4
languages) and for R, I would appreciate learning how
other users who do the same set up their directories
to facilitate distributing libraries for both R and
S-Plus in a way that is also consistent with CVS.  In
other words, I would like to be able to construct
distributions suitable for CRAN without making it too
hard to run Unix/Linux shell scripts that also create
S-Plus distributions.  I am using a single code base
for all platforms except for Fortran and C code
where I allow for single precision objects in S-Plus.

As I am now Microsoft-free (one of my colleagues creates
Microsoft Windows files for S-Plus; I have found that
Linux is amazingly more reliable and powerful than
Windows.  I have set up a web page at
http://hesweb1.med.virginia.edu/biostat/s/linux.setup.html
to help Windows users set up Linux) I am honestly not 
motivated to do special work to make a Windows distribution
for R.  If a Windows R user who has suitable Fortran and
C compilers wishes to periodically create a Windows
distribution once the libraries work under R on
Linux/Unix that would be great.

I am very glad to join the talented R community, and look
forward to collaborating with other users.

Sincerely,

Frank Harrell