[R] Popularity of R, SAS, SPSS, Stata...

[Daniel Nordlund]

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
> On Behalf Of Muenchen, Robert A (Bob)
> Sent: Friday, June 25, 2010 5:39 AM
> To: Liviu Andronic
> Cc: r-help at r-project.org
> Subject: Re: [R] Popularity of R, SAS, SPSS, Stata...
> 
> 
> 
> >-----Original Message-----
> >From: Liviu Andronic [mailto:landronimirc at gmail.com]
> >Sent: Friday, June 25, 2010 7:15 AM
> >To: Muenchen, Robert A (Bob)
> >Cc: r-help at r-project.org
> >Subject: Re: [R] Popularity of R, SAS, SPSS, Stata...
> >
> >On Sun, Jun 20, 2010 at 2:31 PM, Muenchen, Robert A (Bob)
> ><muenchen at utk.edu> wrote:
> >> come up with so far at http://r4stats.com/popularity . I'm sure people
> >> will have plenty of ideas on how to improve this, so please let me
> >know
> >> what you think.
> >>
> >This is not much of a metric, probably not even a ballpark, but I have
> >a habit of measuring the popularity of a software by the number of
> >unread messages in my mail account, sent to one of its main mailing
> >lists. For example, I subscribed to Gentoo, Xfce and LyX MLs much
> >earlier than to that of R, but R quickly and surpassed all in number
> >of unread messages. At the moment I have the following: R ( 37k), LyX
> >(10k), Debian (7k), Xfce (<3k), Geany (.5k). I dare say that R might
> >be more popular than Debian, but again, any such estimation seems
> >farfetched.
> >
> >Regards
> >Liviu
> 
> Hi Liviu,
> 
> E-mail was the thing that got me back to this paper. I had been working on
> variations of measures for several years & was frustrated mostly by how
> many problems I ran into regarding search logic ("SAS" stands for about 15
> scientific topics and of course "R" is far worse). I have all my listserv
> email routed to a set of folders which I always empty at the same time. I
> noticed that recently R-Help had really taken off and that Statalist had
> surpassed SAS-L. So I got the latest monthly data from the listservs and
> switched the program from doing yearly counts to means of the monthly
> figures so I could add 2010 to it. Figure 1 at
> http://r4stats.com/popularity is indeed the number of emails send by each
> of the listservs. All these measures have their own limitations, but I
> find that graph the most interesting since it includes the trends across
> time.
> 

Looking at the help mailing lists has pluses and minuses.  The more popular a software system, the more one might expect an increase in mailing list traffic.  However, that is offset by how easy the software is to use, and how readily one can access other help (written documentation, publishers support sites, user group papers, books, etc.).  

In the case of the SAS-L mailing list, there is one more complication.    Over the last several months, the link between the mailing list and the comp.soft-sys.sas newsgroup has been broken, so if one just looks at the SAS-L traffic it is down substantially because the newsgroup posts are no longer coming through the gateway.  Just one more glitch that makes it difficult to compare the popularity of different systems.

Dan

Daniel Nordlund
Bothell, WA USA