[O] Discovering Babel and R combination!

[François Pinard]

Hi, Org people.

A while ago, I started to get acquainted a bit with Babel, mainly as a
curiosity for an area of Org I did not know nor use yet.  However, for a
couple of weeks now, I'm starting to better understand how useful the
combination of Babel and R may be in real situations.

While documenting a few databases, exploring them with R mainly, I first
developed the habit of saving some of my R code aside, as a reminder to
myself for later retries or further discussions with colleagues.  But
I'm finding out that it is much more fruitful to put the R code directly
in my documentation, taking advantage of Babel to insert the R output
and graphics right into it.  It gets especially interesting because of R
sessions within Org, which may be used to cache prior computation
results, yielding comfortable interaction speed.  Also, despite hundreds
of included graphics already, I do not feel any slowdown in Emacs yet.

I think I'm going to just love it! :-)  Thanks to all contributors!

------------------

There are tiny improvements or questions which I would like to discuss,
however.  As this is all new and a bit overwhelming for me, I apologize
if my reports are a bit fuzzy, I guess dust should settle as days pass,
and would likely develop a clearer understanding with time.

One thing is that I very often have to do "C-c C-x C-v C-c C-x C-v",
that is that I toggle in-lining of images out and in.  It seems that
whenever I save the file, or use "C-c C-c" within an R Babel block, the
images stop being in-lined, while the in-line image flag is not reset.
Ideally, saving a file should not hide in-lined images.  If, for some
technical reason, this is unavoidable, then at least, the in-line image
flag should always tell the truth, so "C-c C-x C-v" would be sufficient
to recall back the images.

Real fun would be that any "C-c C-c" which triggers the re-computation
of an already displayed graphics, merely gets the displayed graphics to
get updated in place, without any more special interaction needed to
re-in-line it.

Another point which gave me some fight to do is the disappearing of
column and row headers in non-graphical output.  I only get the raw data
of the results.  I found it quite annoying with R table() output, for
example, which are rather meaningless with no titles at all.  Currently,
the only reasonable solution I have is to use ":results output org",
combined with an ascii() call on the R side, once ascii configured to
create Org style output (quite a nice feature, should I say!).  Yet, it
would be all nicer and cleaner if none of this extra machinery was
required.

A very minor point is that, within a #+BEGIN_ORG / #+END_ORG block, I
sometimes see an extraneous space at the very beginning, before the
first "|".  Sometimes I do not see it, and output is perfect.  I'll try
to find some coincidence with other things, that would allow to
hypothesize a cause.  Or else, to reproduce the problem with data which
is public enough that I could share it.

I call it a day for now and get some sleep :-).

With enough luck, will find some time to revisit this tomorrow.  Yet
before leaving today, I wanted to say and share my enthusiasm for this
Org Babel / R combination, and also report the tiny problems above.

Keep happy, all!

François