[R] R reports

[Barry Rowlingson]

On Sat, Aug 21, 2010 at 4:09 AM, Donald Paul Winston
<satchwinston at yahoo.com> wrote:
> The ability to generate standard detail, summary, cross-tabs, and control break reports is very important in government and corporate enterprises.

 The great thing about standards, as a wise man once said, is that
there are so many of them. I'll point out that he was speaking
ironically.

 Your standard reports are not my standard reports. I want my heading
typeset in Futura, and you want them in Comic Sans. I want my plots as
barcharts, you want pies. The 'geeky' programmer mentality ingrained
in R means that you can have that, but it won't be given to you.
Here's the tools. Build it, build it *once*, build it good, and then
use it as much as you like.

 The problem with standardized reports is that people's eyes glaze
over. Ever seen a corporate report that was sixty almost identical
report pages - graph here, table here, pie charts here? You stop
looking after the third one. Of course government and corporations
must be doing something right because they're just doing so well at
the moment. Oh wait, maybe not. Meanwhile UK universities are
currently trying to stuff more new students into oversubscribed
courses than ever before. Who trusts government and corporates these
days? Don't put them up as shining examples of the way to do things.
It's revolution (with a small r)  time, and I would hope that academia
has a part to play.

> The people who develop R and who use it live in a small ghetto and don't get out much.

 Please please please come along to the useR meeting next year in the
UK - its the one time of the year when the R developers are let out of
their ivory towers for a week to mingle with the unwashed masses.



Barry