[R] The Origins of R

[John Maindonald]

In another thread on this list, various wild allegations have been  
made, relating to the New York Times article on R.  I object both to  
the subject line and to the content of several of the messages, and  
will not repeat or quote any of that content.  It smacks to me of  
mischief making.

Discussion has centered around the following quote from the NY Times  
article:
“According to them, the notion of devising something like R sprang up  
during a hallway conversation. They both wanted technology better  
suited for their statistics students, who needed to analyze data and  
produce graphical models of the information. Most comparable software  
had been designed by computer scientists and proved hard to use.”
The comment that "the notion of devising something like R sprang up  
during a hallway conversation" is strictly true.  Certainly, this  
seems like a very plausible account.  I'd have more difficulty  
believing that the notion was communicated to them in separate  
dreams.  Part of the wanted technology was freedom for students to  
take the software home, or copy it down from the web.
There was a further story to be told, about the origins of the  
language that Ross and Robert implemented and adapted.  The NY writer  
pretty much left out that part of the story (S did get a mention, but  
its connection with R did not), but did remedy this omission in a  
follow-up.
Nor did the article do much to acknowledge the workers and work that  
has gone into R's continuing development. Getting the attributions  
"right" is difficult.  Even if "right" according to common conventions  
(and one can argue as to just what the conventions are, especially in  
the matter of computer language development), they are unlikely to be  
totally fair.  Stigler's Law of Eponomy has wide sway!

In the preface to the first and second edition of "Data Analysis and  
Graphics Using R", we have:
"The R system implements a dialect of the S language that was  
developed at AT&T Bell Laboratories by Rick Becker, John Chambers and  
Allan Wilks".
The only 1st edition attribution to Ihaka and Gentleman was in Chapter  
12: "For citing R in a publication, use Ihaka and Gentleman (1996)".   
[NB: Type citation() to see the form of citation that should now be  
used.]
That was as it now strikes me unfair to Ross and Robert, but no-one  
complained.  Perhaps no-one ever read that far through the preface!

There's an excellent brief summary of the history of R, and its  
connections with S, in Section 1.4 of John Chambers' "Software for  
Data Analysis".    Appendix A has further details on the development  
of S, a kind of pre-history of R.

John Maindonald             email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473    fax  : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.