Ok, I am giving it a shot. Here is my setup:

;;;;;;; org path
(defvar org-load-path
  (list (file-name-as-directory
 (expand-file-name "org" starter-kit-dir)))
  "List of directories to find org-files that
  `org-babel-load-file' can load code from")

(defun org-require (feature)
  "Load a FEATURE from an org-file.
FEATURE is a symbol, and it is loaded from an org-file by the name of
FEATURE.org, that is in the `org-load-path'. The FEATURE is loaded from
`org-babel-load-file'."
  (let ((org-file (concat (symbol-name feature) ".org"))
(path))

    ;; find the org-file
    (catch 'result
      (loop for dir in org-load-path do
    (when (file-exists-p
   (setq path
 (expand-file-name
  org-file
  dir)))
      (throw 'result path))))
    (let ((default-directory (file-name-directory path)))
      (org-babel-load-file path))))

(org-require 'org-ref)
(org-require 'doi-utils)
(org-require 'org-show)


In the only directory I define in org-load-path, I put org-show.org,
org-ref.org, and doi-utils.org, and the code above is in an init file. When
I load emacs, for these files, so far, it works great. They are simple
files that only tangle to one .el file so far.



John

-----------------------------------
John Kitchin
Professor
Doherty Hall A207F
Department of Chemical Engineering
Carnegie Mellon University
Pittsburgh, PA 15213
412-268-7803
http://kitchingroup.cheme.cmu.edu



On Sun, Jul 27, 2014 at 5:50 PM, Bastien <bzg@gnu.org> wrote:

> Hi John and Rainer,
>
> Rainer M Krug <Rainer@krugs.de> writes:
>
> >> What if we created a new directory in the repository called "org" which
> >> contains these kinds of files?  It would be analogous to the "lisp"
> >> directory. I don't think we need to have both ob-R.org and ob-R.el in
> the
> >> repository.
> >
> > I think that would be a very good idea for certain modules like your
> > org-ref and ob-R I am working on.
>
> Maybe you could use such a "load-flow" for a while, see if it is
> robust enough, and if so, we could move on and have it in the repo?
>
> My gut feeling is that it will bring some issues, so experimenting
> before would be nice.
>
> --
>  Bastien
>