Re: [Orgmode] GTD: Solving the bottleneck of converting incoming mails to NEXT actions

[Adam Spiers]

On Mon, Sep 24, 2007 at 10:44:44PM +0200, Georg C. F. Greve wrote:
> Hi all,
> 
> I've now been using Org mode for a couple of months and am quite happy
> with it in most ways, with conversion of incoming emails into NEXT
> actions being the main bottleneck.

[snipped]

> Since I know that various other people are using Org mode for GTD kind
> of methodologies, as also discussed in
> 
>  http://thread.gmane.org/gmane.emacs.orgmode/523
> 
> I wonder whether anyone solved this issue already.

Yes, I used to have this very same bottleneck.  It's not perfect yet,
but a lot better than it used to be.

I posted details in a previous thread on this list about how I use a
combination of mutt, mairix, shell wrappers, and elisp to extract a
mairix:$message_id style hyperlink from the currently viewed mail
within mutt, and then yank it into my TODO.org emacs buffer, manually
creating structure around it (I almost certainly need to use
remember.el to streamline things further).  Here's an example:

  * PROJECT [#B] submit newer expenses          :@internet:admin:
  ** NEXT submit to iExpense                    :sub60:
  *** don't forget [0/3]
     - [ ] HP ProCurve switch
     - [ ] linuxconf.eu
     - [ ] Travelodge Birmingham Maypole
           e-receipt
           <mairix://m:address@hidden>
  ** NEXT read mail from Donna about where to send :sub10:
     <mairix://m:address@hidden>

You can see two hyperlinks in there (yes, I still need to tweak my
code to take advantage of the functionality recently added in newer
versions of org.el) - with a quick keystroke I can pop up an xterm
with mutt viewing the relevant mails.  They're referenced by message
ID, so I can archive them in any folder, and mairix will still
retrieve them instantly.

There's a more general point worth making about the indexed search
capability which mairix gives me: I no longer need to archive by topic
because I can always retrieve instantly via a search.  Thus these days
I archive by date into monthly buckets, and mutt lets me do that with
two keystrokes, e.g.

  # Hairy mutt configuration to automatically archive into monthly
  # buckets based on date in mail header.  The trickery with the 15th
  # of the month is necessary because date -d 'last month' doesn't
  # work at the end of long months - see date(1) info pages.
  fcc-save-hook ~r\ `date +'1/%m/%Y-31/%m/%Y' -d "$(date +%Y-%m-15) -2 months"` 
=archive/`date +'%Y-%m' -d "$(date +%Y-%m-15) -2 months"`
  fcc-save-hook ~r\ `date +'1/%m/%Y-31/%m/%Y' -d "$(date +%Y-%m-15) -1 month "` 
=archive/`date +'%Y-%m' -d "$(date +%Y-%m-15) -1 month"`
  fcc-save-hook ~r\ `date +'1/%m/%Y-31/%m/%Y' ` =archive/`date +'%Y-%m'`

This is a triple win:

  - Often archiving by topic didn't make sense anyway because many
    mails cover multiple topics.

  - Archiving by topic substantially increased the mental overhead in
    processing a single mail, thereby exacerbating the bottleneck you
    refer to.

  - The mail can vanish from my inbox the moment it has been
    "processed", even if I haven't completed the resulting action.
    This retains (in theory!) the "hard edges" between unprocessed
    and processed buckets, which means I can realistically aim to have
    an empty inbox very frequently.  Ahhh, sweet relief :-)

What I personally struggle with most is regularly reviewing TODO.org -
but that's another story ...