Org-publish: adding a new format?

[Bill Powell <bill@billpowellisalive.com>]

Hi all,

Does anyone have any suggestions for adding a new format to
"org-publish"? Any articles, or threads on the mailing list? I've
looked around but haven't found anywhere to start.

I want to convert an org file to a tab-separated format, so that it
can be imported into a flashcard program called Anki
[[http://ichi2.net/anki/]]. The basic idea is that each *** header is
a prompt, and the following text is the answer.

*** Who wrote /War and Peace/?

Leo Tolstoy.

*** What is the preferred editor for using Org-mode?

Windows Notepad.


and so on.

The catch is, I'd also like to preserve the *bold* and /italic/ formatting.
Not to mention those awesome org-tables.

Now, Anki understands HTML formatting. So, ideally, I can use the
excellent export to HTML that org-publish already does. That's 95% of
the work right there.

In fact, I have already written a Perl script that takes an exported
HTML file and chomps it into a .tsv. (Almost.)

But although my elisp is pretty rudimentary, I have a couple reasons
I'd like to implement this within org-publish instead:

1) I could share my work (at least by posting it on my web site; my elisp
isn't ready for official contributions yet), and

2) It might make it easier to access the different org-elements
directly. For instance, I use *** headers as prompts, so that I can
use * and ** to organize the flashcards in the org-file. But, I would
like to be able to save the * and ** headers, and use them as "tags" in
the final Anki flashcard. (Anki lets you tag your flashcard.)
Similarly, I'd like to convert org-mode :tags: as well.

I can do all this in Perl, but it feels messy and inelegant, not to
mention brittle.

On the other hand, those who know org-publish may feel that this is a
tall order for org-export-html-final-hook, and just the sort of job
that Perl does best.

Any thoughts?

Thanks,
Bill Powell