[R] Wikis for R

[Drew]

Frank uses the term "hierarchical keyword
organization" which I agree is a good way to organize
a system designed to help users. In fact, this is one
reason why I like the R graphics gallery which allows
one to quickly find a particular type of plot based on
keyword, examine the plot to see if it's close to
what's desired, and then get the detailed code to
examine or modify for one's own specific purpose. It's
an example of going from general concept to specific
details. 

R's HTML help pages already have a keyword system but
it takes you to the specifics too quickly. I'd
recommend revising the current keyword system in the
help pages (or adding to it) to create what Frank
calls "data manipulation examples gallery." In other
words a code gallery that users could browse quickly
and simply copy and paste code from it to their
application. Think of the keywords as different "How
to" topics (e.g. how to create a data frame, how to
reshape data, etc.). 

** I think this sort of thing would save time for
users of all levels.**  New users could find code
quickly WITHOUT submitting a request to R-HELP email
list. Advanced, experts, and gurus would no longer
have to respond to such requests for the umpteenth
time. Think of this as the examples that did not make
it to the help files. Because it's a wiki folks could
add to it the examples that are truly helpful.

Creating such a system requires 3 steps (at a very
high level):

1) ** Develop a hierarchical keyword system **
Combine the current keyword system
(http://finzi.psych.upenn.edu/R/doc/html/search/SearchEngine.html)
with something more like the left-hand frame of the
function finder
(http://biostat.mc.vanderbilt.edu/s/finder/finder.html).
Essentially a list of keywords organized by "how to do
X in R". Users could browse this list and click on a
"how to" topic to get to more specific keyworks or see
the code gallery for that particular keyword. Once
this list of keywords gets generated then it would
would remain mostly static over time. If changes are
required then have the site maintainer add the word,
but the keyword list is not open to public revision.


2)  ** Add code as content for each keyword. **
Without quality and quantity of content, this proposed
system is not worth much. No one will use it without
content -- I don't care what fancy tool we use to
create or manage such a thing. I suspect that much of
the code to start this gallery could come from R tips
(http://www.ku.edu/~pauljohn/R/Rtips.html). After
amassing a good size collection of code, allow the
public to start adding to the code (but not the
keywords identified in #1 above.)


3)  ** Link 1 and 2 above, which are on the wiki, with
the existing help files that get installed with R **. 
So when a user chooses "HTML Help" from the Help menu
(in the windows version of R, for example) he/she
would see an additional clickable link that says
something like "How To". Clicking on this would take
the user to the list of keywords for the code gallery
on the online wiki. This is their starting point. From
here the user would click on appropriate keywords to
browse the code gallery. The code itself could be
clickable to bring up the very detailed help pages
much like the R Graphics Gallery does (for example
clicking on the green color "rnorm" in the first graph
in the Graphics Gallery takes me to the following
hyperlinked help page:
http://addictedtor.free.fr/graphiques/help/index.html?pack=stats&alias=rnorm&fun=Normal.html).


I'm wondering how much of the above could be done
without requiring much work from the R CORE team? I'd
recommend a small group of individuals (10 - 20 or so)
to start getting the content together since that may
be the most time intensive part. (You can't expect 1
person to do it all, and 1000's of people won't make
much progress.) I can lend a few hours per month if
there was a cadre of dedicated individuals who are
passionate about making this happen. However, I have
very few programming skills, almost none outside of R,
of which I'm a relative beginner.

This is the vision I have for the wiki. What do others
think?

~Nick


-----Original Message-----
From: r-help-bounces at stat.math.ethz.ch
[mailto:r-help-bounces at stat.math.ethz.ch]On Behalf Of
Frank E Harrell Jr
Sent: Saturday, January 07, 2006 6:05 AM
To: Detlef Steuer
Cc: r-help at stat.math.ethz.ch; phgrosjean at sciviews.org;
steuer at unibw-hamburg.de
Subject: Re: [R] Wikis for R


Detlef Steuer wrote:
> On Thu, 5 Jan 2006 11:23:04 -0200
> "Fernando Henrique Ferraz P. da Rosa"
<academic at feferraz.net> wrote:
> 
> 
>>Martin Maechler writes:
>>
>>> If you go to the bottom of that wikipedia page,
>>> you see that there is an "R Wiki" -- and has been
for several
>>> years now (!) at a Hamburg (De) university.
>>>
http://fawn.unibw-hamburg.de/cgi-bin/Rwiki.pl?RwikiHome
>>>
>>>(...) 
>>>So, are you sure that another R Wiki is desirable,
rather than
>>>have people who "believe in Wiki's for R" use the
existing
>>>one(s)?   I believe the main challenge will
(similar as for
>>>an "R-beginners" mailing list) to have
well-qualified "editors"
>>>to be willing to review and amend what others have
written.
>>
>>        IÂ´ve tried to colaborate on the R Wiki
hosted by the Hamburg
>>university but the Wiki would get regularlly
vandalized by some spam
>>bot, and then I'd have to manually keep reverting it
several times. Also
>>the wiki engine used by this wiki is very
rudimentary. I think the
>>DokuWiki engine, which is used by Philippe Grosjean
is more promising as
>>a workhorse for an 'official' R-wiki. 
> 
> 
> What Fernando says is mostly true. I installed the
UseMod Wiki after talking to some of "us" as a test
balloon. The engine was chosen according to simplistic
installation and a "just-enough" feature set. Perhaps
that was a wrong decision. I`m one of the believers in
wikis, but mine over here did not fly.
> 
> Probably this and any wiki needs a critical mass for
visitors beginning to return and collaborate and
therein the Hamburg wiki failed. (One day I wanted to
take down the wiki for apparently having no users, but
just then someone gave some feedback)
> 
> As Frank Harrel pointed out, this may be because
R-help is just too helpful.
> In contrast to Frank I don`t think we should abandon
e-mail because of its success. The mailing list is, in
my opinion, the single biggest plus R has above all
competition.
> The wiki should provide something complementary to
r-help. Btw. those who do not search for information
in the mailing list archives or on CRAN before asking
simple questions won't do so in a wiki or in a
bulletin board system. (I find those simply unusable.
Am I getting old? I want to edit using my favourite
editor, not in some browser window.)
>  
> A central place for example code was my intention,
when opening the wiki back then. 
> 
> Back to operating wikis:
> The wiki spamming is a serious problem, especially
because I HATE to login to read or edit anything. So
the choice is: take the wiki as seriously as work and
have a look every other day to remove the spam (or
better: form a group of volunteers). That hurts or at
least is no fun. Or put restrictions on it. That hurts
even more. Perhaps I do not understand Philippe`s
"loggable". What does a logfile with IPs help? The
spammers are strangers selling ******; I don`t want to
find them :-)
> 
> To sum it up:
> There is a very simple way to proceed:
> Philippe uses his Docuwiki install  as official,
_general_ Rwiki and I close down mine. The beginners
will find their niche in there, if there is a real
demand. 
> I wouldnÂ´t mind to give up "my" wiki, because I
have to admit it failed to achieve what I would have
liked.
> 
> So, Philippe, if you like, you can take over. I
would replace my wiki with a notice where to find
yours and the community gets a second chance :-)
> 
> Detlef

The current e-mail system places a low burden on users
if they follow 
basic posting rules.  The burden is too low and users
still do not 
search for past answers and we also get dozens of
separate messages on a 
single topic (e.g., ylim on barplots).  As long as
everyone allows this, 
a wiki or discussion board will not work.  We need to
rethink the e-mail 
system in my view to create motivations for approaches
with true 
memories and hierarchical keyword organization instead
of using the 
apparent memory-less system.  This should be thought
through to include 
the new graphics gallery and a data manipulation
examples gallery, and 
other things, and needs to be launched from
www.r-project.org IMHO.

Frank

> 
> 
> 
> 
>>        I think that the title could be perhaps
changed to Rwiki
>>and the contents currently hosted on the Hamburg
wiki 'transfered' to 
>>the new location, if the current mantainers of the
Hamburg Wiki and
>>Philippe Grosjean agree (IÂ´m cc-ing this msg to
them).
>>
>>        This could emerge then as official or
semi-oficial R-wiki, to be
>>linked to from the R-project home. 
>>
>>
>>--
>>"Though this be randomness, yet there is structure
in't."
>>                                           Rosa,
F.H.F.P
>>
>>Instituto de MatemÃ¡tica e EstatÃ­stica
>>Universidade de SÃ£o Paulo
>>Fernando Henrique Ferraz P. da Rosa
>>http://www.feferraz.net
>>
> 
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html


-- 
Frank E Harrell Jr   Professor and Chair          
School of Medicine
                      Department of Biostatistics  
Vanderbilt University

______________________________________________
R-help at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html