This paper inspired me to check out Emac's org-mode a bit and currently I try to assess what's more suitable for writing my documents: knitr/Sweave (I'm mainly using R to do my programming) or org-mode.
What I really like about knitr is the option to externalize the actual source (watch out: the declaration of labels/names in the R script seems to have changed from ## ---- label ------- to ## @knitr label; see ?read_chunk) and "link" them to the actual literate programming/reproducible research document (as opposed to actually writing the code in that very document):
"Import" with
<<import-external, cache=FALSE>>=
read_chunk('foo-bar.R') # has label/name 'foo-bar'
@
and "re-use" by referencing the respective labels with
<<foo-bar>>=
@
Is this also possible in org-mode or am I bound to putting the actual code into the .org document?
I found this, but I did not find any specific notion of linking/importing external source code files and be able to execute them by having the linked code inside
#+BEGIN_SRC R
<linked code>
#+END_SRC
I do see that this approach might contrast the general paradigm of literate programing to some extend. But I like to work in a somewhat "atomic" style and thus it feels more natural to me to keep the files separated at first and then mash everything together dynamically.
Would named code blocks help?
#+NAME: foo-bar
#+BEGIN_SRC R
source(foo-bar.R)
#+END_SRC
And then evaluate (i.e. load) the code when you actually need it:
#+CALL: foo-bar()
See manual for more details.