A couple of months ago, I asked Frank Harrell (fharrell@virginia.edu), the
author of the Hmisc S-Plus library, about the reason for his "No porting to
R" requirement. I may have previously posted his reply, but I don't
see it
in the archives, so I'll send it again.
This is a edited version of our email conversation:
Gregory R. Warnes:
The license to the Hmisc library explicitly prohibits modifying the
code to work with "R". Why have you made this restriction?
Frank E. Harrell, Jr.:
Some time ago an R user took it on himself to start converting my
libraries to R without at least sending me an E-mail. He started
the conversion at a time just when I was making a good deal of
error corrections and extensions to one of the libraries, so that
the R version would be instantly out of date and buggy. I thought
this was strange and wanted to prevent such problems in the future.
In the meantime I have been looking at R a good deal to see what
would be involved in conversion. I have found many more
inconsistencies between R and S than I originally thought. Some of
these inconsistencies are because of improvements, some make the
language worse (e.g., omission of single precision objects in R, the
inability to attach directories). I estimate it will take me about 200
hours of work to convert, even if I didn't mind losing single precision
and Trellis graphics in R. I'm not ruling out a conversion in the
future
but it will take some resources.
If resources become available in the future I would be interested in
starting a CVS server to work will a good R programmer in this
endeavor.
Above all I do not want two code bases; version control is needed with
a
single code base.
-Greg
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
