[R] Size of R user base

[Martin Maechler]

>>>>> "BXC" == BXC (Bendix Carstensen) <bxc at steno.dk>
>>>>>     on Tue, 20 Apr 2004 09:10:42 +0200 writes:

    BXC> I wonder if there is a count of the number of downloads
    BXC> of each version of the R installations (for example
    BXC> rw1090.exe)?

doesn't work either for several reasons,
some being

- students get R (and ESS/WinEdt and ...) on CDs
  or download from local repositories

- many Unix users in the same organization use R installed only
  once in a mounted (aka "network") file system. 
  Windows users as well, e.g. here at ETH, and probably in many
  other places.

----

In general, I quite agree with Uwe's comments --- including those
about the ratio of R users and R mailing list subscribers; I'd
guess that to be around 10:1, though rather higher than lower.

About a `confidence interval': If we log transform it becomes
feasible to give an interval I'm ``quite confident'' about:

  log10(N) \in [3.5, 5.3]

but then that's probably not informative enough for management
decisions.

Martin