[Orgmode] Re: [Babel] now understands org-mode lists

[Eric Schulte]

Hi Jambunathan,

Jambunathan K <address@hidden> writes:

[...]
>
> To summarize, what I am really thinking is this:
>
> 1. Use C-c C-c (org-dwim) with a prefix modifier key that makes it
>    (org-custom-dwim). 
> 2. The custom dwim pops up an easy menu that is context dependent. For
>    example, for the list case it could offer a choice of numbering
>    styles - alphabetical, subsection etc.
> 3. Make Babel recognize implicit arguments and results. So when on a
>    Orgmode list (or when the active region is a Ormode list), Babel
>    creates a virtual 'input' var that represents the textual list and
>    does 'replace-region' for the results block. So there is no more
>    '#+results' that the user sees in the text document
> 4. The easy menu keys in (2) maps to 'library of babel' calls that
>    behave as in (3) which creates needed overlays.
> 5. Make the exporters (specifically the ASCII exporter) to use overlays
>    to emit the custom-styled lists.
>
> The above idea is more or less along the lines that I exchanges with you
> a few months back where I (tried) articulating the idea of Babel as 'a
> text manipulator' of Orgmode entities like headlines etc.
>

Interesting, a couple of thoughts...

- I think that your idea of a quick access menu for library of babel
  calls which converts the objects (table, list, etc...) under the point
  into an implicit argument to the code block could make lob functions
  much more accessible and easily usable.

- I'm not so sure about the use of `replace-region' in (3) as I can
  imagine that there would be cases where you would rather perform some
  operation on the input data for pasting elsewhere in the document.

- I can see how this would quickly lead to a desire for type hinting on
  library of babel code blocks, e.g. if my point in on a table I don't
  want to see all of this LOB functions which operate on lists.  such
  type hinting could easily be implemented as another header argument

- This would certainly be helped by the ability to pass un-named
  arguments to code blocks (something I've been meaning to implement
  anyways)

I agree this is an interesting way to use Babel.  Unfortunately I simply
don't have the time to develop this myself, however if you want to start
developing in this direction I'm happy to help in terms of answering
questions and adding features on the babel side of the fence.

Best -- Eric