[Rd] Help Documentation

[Henrik Bengtsson]

Dear all, 

without taking sides here, I see two major advantage of keeping the
redundancy in any documentation minimal. First, it makes the
maintanance of the documentation as simple as possible. This in turn,
minimizes the risk for getting inconsistent documentation in new
updates. Otherwise, someone has to have a really good overview and
know where to update when, say, one default argument is updated (or we
have to live with incorrect documentation).

One possible solution to a documentation for beginners is to have a
separate package just for the documentation. In that package you can
document ?exit etc . Load the package and help.search("exit") will
find "anything" regard exitting. To get started with this you have to
know how to write Rd documentation (very similar to LaTeX). You'll
find details in the help; type help.start(). 

Cheers

Henrik

> -----Original Message-----
> From: r-devel-bounces at stat.math.ethz.ch 
> [mailto:r-devel-bounces at stat.math.ethz.ch] On Behalf Of Gabor 
> Grothendieck
> Sent: den 29 mars 2004 01:44
> To: ivo.welch at yale.edu
> Cc: R-bugs at biostat.ku.dk; r-devel at stat.math.ethz.ch
> Subject: RE: [Rd] Help Documentation
> 
> 
> 
> 
> I think many people share your view and are aghast at the 
> reception that some well-intentioned posts receive.  There 
> have been past discussions on this and many people feel the
> way you and I do.  
> 
> Just to head off another round, let me acknowledge that 
> there appears to be multiple viewpoints and although hard 
> to believe by myself, there actually is a contingent that 
> views what I see as insulting responses as appropriate.
> 
> ---
> From:   ivo welch <ivo.welch at yale.edu>
> 
> ladies and gents:
> 
> I have posted a couple of simple questions recently. As often
happens 
> to novices, the information was there somewhere, even in front of my

> eyes, and I just did not see it. I looked in docs that seemed 
> to me to 
> be the right place for this particular information, but did 
> not find it. 
> There is no question: mea culpa, and everything is documented 
> somewhere 
> in R. (Worst comes to worst, it is documented in the source.)
> 
> But here comes my complaint: I tried to help by documenting 
> where I got 
> lost, and by suggesting simple one-liners for the 
> documentation, which 
> would provide additional cross-references to what I was looking for.

> The cost of adding additional brief sentences to the help must be 
> relatively small, and the help to stuck novices may be 
> considerable in 
> reducing the learning curve. For my specific examples, I suggested a

> reference to q() in ?exit, and a "select= -c(v1,v2)" to ?subset.
> 
> Clearly, the information is redundant. (Of course, in a sense, all 
> documentation is redundant.) The goal of good documentation should
be 
> to help novice users who do not know the answer. The goal 
> should not be 
> minimum redundancy in the help files. Being fairly new to R, I see 
> difficulties where Brian Ripley and other experts and developers no 
> longer do. I bet that if I wonder about the answers, I am more than 
> likely not alone. In fact, I think it would really make sense to 
> improve the docs by studying where novices get stuck.
> 
> I was told by Brian to stop sending such suggestions, in order not
to 
> clutter the R bug report list. OK, I can save my time; I just 
> wanted to 
> help. But, for others' sake, please reconsider the policy of not 
> gearing the internal R documentation for novices like myself.
> 
> I will butt out here.
> 
> regards,
> 
> /ivo
> 
> PS: Incidentally, the R help seems a little schizophrenic. For 
> example, Brian Ripley is the most helpful source for learning R
(both 
> books and posts), and I am rather grateful for it. I just do not 
> understand why, at the same time, he seems to be annoyed 
> while fielding 
> questions of the r-help post-list. He is not the only individual who

> likes to help, but grudgingly so...
> 
> ______________________________________________
> R-devel at stat.math.ethz.ch mailing list 
> https://www.stat.math.ethz.ch/mailma> n/listinfo/r-devel
> 
>