[R] two questions for R beginners

[Paul Hiemstra]

Jack Siegrist wrote:
> My biggest impediment, as a scientist without previous programming
> experience, is that the R help is not beginner-friendly. I think it is
> probably great for experienced programmers and for the people who helped to
> create the software, to help them  remember what they did, but I think it is
> very difficult for a newcomer without a strong programming background to
> learn about a new function or to discover the name of a function that you
> are pretty sure should already exist. Maybe this wouldn’t matter for most
> programming languages, but as free statistics software R is obviously going
> to attract many scientists who want to get an analysis done and have varying
> levels of experience with programming. 
>   
Hi Jack,

A problem more or less is that the R community consists primarily of 
volunteers. People who answer questions to the help list in their spare 
time or during company time. This also holds for many of the online 
material. A program like Mathematica has a company providing the online 
material, they hire people to do this work. I don't use this as an 
excuse for R, but it might explain why the R community is what it is.

In reply to the 'bashing' of new users. I agree that sometimes the 
experts answering the questions can be blunt, but most often it is in 
response to questions that are very hard to answer. As I said earlier in 
this mail thread, asking the right question already involves some of the 
knowledge to answer the question. So to get good, informative responses 
a user needs some level already.

I do want to point out that there is a posting guide for the mailing 
list that gives a quite detailed instructions, like give the exact error 
(don't just say, R crashes). Provide traceback() and sessionInfo() etc, 
etc. And a lot of posters do not adhere to these rules.

cheers,
Paul

> I found it much easier to learn how to use Mathematica, using only the
> online help. With R I had to buy several books to get a handle on it, which
> is fine, but even the books that I have found to be most useful tend to be
> didactically lacking—either too cursory or mired in unexplained programming
> jargon. They are OK just not great.
>
> What I think would be very helpful is an introduction to programming using
> R, preferably a big thick college textbook that takes at least a semester to
> go through, which should be a prerequisite for going through the
> Introduction to R available on CRAN.
>
> Also to do any analysis on real data you have to use the apply family of
> functions to perform different functions by groups. A long introduction to
> these functions, with lots of comparisons and contrasts between them would
> be very helpful.
>
> A few random examples concerning the R help: 
>
> In my version of R (2.7.0 on Windows XP) typing
>   
>> ?+
>>     
> doesn’t do anything, but then if you type in the next line
> + ?sum
> you get the “Arithmetic Operators” help page.
> If you had just typed
>   
>> ?sum
>>     
> in the first place you get the “Sum of Vector Elements” help page. 
>
> Most examples in the R help pages use way to many other functions to be
> useful to a beginner. If an example uses 10 other functions besides the one
> being described, chances are a beginner won’t know what one of them does,
> which can set off a chain of having to look up other irrelevant functions.
>
> Some function names in the base package are goofy, such as “rowsum” which is
> used to “compute column sums across rows”, not to be confused with “rowSums”
> which computes row sums.
>
>   


-- 
Drs. Paul Hiemstra
Department of Physical Geography
Faculty of Geosciences
University of Utrecht
Heidelberglaan 2
P.O. Box 80.115
3508 TC Utrecht
Phone:  +3130 274 3113 Mon-Tue
Phone:  +3130 253 5773 Wed-Fri
http://intamap.geo.uu.nl/~paul