[R] mailing list for basic questions - preliminary sum up

Wed Dec 17 09:20:53 CET 2003

On Tue, 16 Dec 2003, Gabor Grothendieck wrote:

> 
> 
> My personal view on this is that there is need for a friendly
> list with a more "customer service" attitude than r-help.

This is not a balanced view. In a project like ours, you really do need to
put participation in balance. If you don't offset perhaps direct but
concise and accurate advice by concurrent commitment to other areas of the
project (often not visible on this list), then you don't get a balanced
picture.

Please also recall that questions asked can be a valuable input to
discussions about when legacy behaviour of functions has become too
difficult to understand - very often leading to changes at least in
documentation. This is a good reason for not trying to divert questions to
a separate list (I think developers and package maintainers would not be
likely to read such a list), lack of shared archives is a second.

What can cause irritation is when list users are given accurate advice,
like read the documentation, read the FAQ, often with hints or actual
solutions, and then come back with the same question, so obviously not
valuing the advice offered. Note that people offering advice try to take
account of language issues, English is not the native language of many
(most?) list users. Also note that the reply time for obvious answers is
very short, most often just the name of the function. Is this impolite,
really?

> 
> r-help is really very useful but its also intimidating
> and I bet lots of people have questions that they never ask
> for fear of the response.   Maybe some of them even decide
> not to learn R.

I do not think you would win your bet. The questions attracting tougher
responses are either the "do my homework for me" type, or the "I'll ignore
the good advice I was given and repeat my question" type. Think of the
list as a graduate seminar - is a sharp comment never appropriate? I'm
sure that I can remember very helpful sharp comments from my teachers and
fellow-students (maybe too few?) that made me see things in a more
appropriate light. One basic characteristic seems to be that if the
question does indicate seriousness about trying to analyse data, respect
for the task at hand, then predictably lots of good advice comes quickly.

I'm also not too sure about the "learning R" question. Of course there is
the GUI/CLI issue, and the "very many defaults already filled in" issue,
but actually market share really isn't a driver here, is it? Isn't this
more about attitude and motivation in taking an active role in analysing
data? If your research question really itches, what should it take to stop
you learning R (or associated packages)? 

As has already been said, there is a lot of documentation. It is possible
that something more like edited weblogs from beginners could be collated
which - if indexed for searching - would function better than a "beginners
list" - a link to something like this was posted a couple of months ago.

> 
> ---
> Date: Wed, 17 Dec 2003 00:49:15 +0100 
> From: Martin Wegmann <wegmann_mailinglist at gmx.net>
> To: Spencer Graves <spencer.graves at pdf.com>, <rossini at u.washington.edu> 
> Cc: <r-help at stat.math.ethz.ch> 
> Subject: Re: [R] mailing list for basic questions - preliminary sum up 
> 
>  
>  
> Hello, 
> 
> I agree completely that well thought out questions are important to receive 
> good and quick replies and I agree as well that the replies on the R-help 
> list are very good and helpful.
> But I had to learn and I am still learing how to write good questions and 
> appreciate Spencer's explanantion how a good question should look like in his 
> opinion. 
> 
> I am not sure how this new mailing list might evolve. 
> It might be that the R-beginner list takes some load of the R-help list by 
> reducing the amount of "basic" questions which won't be questioned anymore 
> here (what aren't many) and that new user might be taught to post "good" 
> question before they start posting to R-help.
> If it proves to be ineffective or might affect R-help in some unwanted manner 
> it would be an easy one to shut it down. 
> 
> I doubt that it will split the R-help list - in my opinion it is unlikely that 
> medium/experienced R user who will subscribe to R-beginner will unsubscribe 
> from the R-help list. 
> Moreover people starting with R are less likely to send any mails to this 
> list, some do and are refered in most cases to the manuals. 
> When I started R I looked through the archive and because I did not understand 
> even one question, I was intimidated by this list and did not send any mail 
> until a few weeks later (that was not because of the statistics but the 
> commands)
> For this kind of people the R-beginner list is thought - to encourage them to 
> send "stupid" questions during their first steps in R. 
> 
> They shall recognize questions they would have asked themselves.
> Therefore I think that the quality of the question is in this case less 
> important than it's level.
> 
> I hope I did not misunderstood some points ,-)
> 
> best regards Martin
> 
> 
> 
> On Tuesday 16 December 2003 17:20, Spencer Graves wrote:
> > I agree with Tony's observation that well thought out questions
> > are more likely to receive an answer than something that is long,
> > rambling, and poorly focused. Many questions take more time to read
> > than I have available, so I don't bother. I like questions that include
> > toy examples in a few lines of code that I can copy from an email into R
> > and test ideas. Careful formatting that looks pretty in an email is an
> > obstacle for me, because it increases the work required to get it into
> > R. Many questioners could answer their own problems in the process of
> > generating such a toy example. When they can't, that exercise helps
> > them focus the question, which makes it easier for potential respondents
> > to understand the problem and reply. Without that, I must either
> > generate a toy example myself (which I've done many times) or respond
> > with untested code and risk looking stupid when my untested suggestion
> > doesn't work.
> >
> > hope this helps.
> > spencer graves
> >
> > A.J. Rossini wrote:
> > >"Pascal A. Niklaus" <Pascal.Niklaus at unibas.ch> writes:
> > >>- In my experience even *very* basic questions *relating to the R
> > >>language* do get answered on r-help. I'm impressed by how much time
> > >>some members of the R core team spend answering relatively basic
> > >>questions, and by how elaborate their answers generally are. So I
> > >>cannot see much need for a new R mailing list. There are these
> > >>excellent mailing list archives, so why "fragment" this list?
> > >
> > >To follow up, well-thought through basic questions do get answered; in
> > >particular, they can be useful for those of us writing packages,
> > >documentation, etc.
> > >
> > >I have a sense that it is the quality of the question (details of what
> > >is intended to do, or not known, signs of using other sources of
> > >materials which folks have spent years on, no signs that this is a "do
> > >my work for me" question) rather than the level of the question, that
> > >is an issue.
> > >
> > >best,
> > >-tony
> >
> > ______________________________________________
> > R-help at stat.math.ethz.ch mailing list
> > https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Breiviksveien 40, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93
e-mail: Roger.Bivand at nhh.no