File: remember.el.html

* The idea

Todo lists, schedules, phone databases... everything we use databases for is really just a way to extend the power of our memory. To be able to remember what our conscious mind may not currently have access to.

There are many different databases out there -- and good ones -- which this mode is not trying to replace. Rather, it's how that data gets there that's the question. Most of the time, we just want to say "Remember so-and-so's phone number, or that I have to
buy dinner for the cats tonight." That's the FACT. How it's
stored is really the computer's problem. But at this point in time, it's most definitely also the user's problem, and sometimes so laboriously so that people just let data slip, rather than expend the effort to record it.

"Remember" is a mode for remembering data. It uses whatever
back-end is appropriate to record and correlate the data, but it's main intention is to allow you to express as _little_ structure as possible up front. If you later want to express more powerful relationships between your data, or state assumptions that were at first too implicit to be recognized, you can "study" the data later and rearrange it. But the initial "just remember this" impulse should be as close to simply throwing the data at Emacs as possible.

* Implementation

Hyperbole, as a data presentation tool, always struck me as being very powerful, but it seemed to require a lot of "front-end" work before that data was really available. The problem with BBDB, or keeping up a Bibl-mode file, is that you have to use different functions to record the data, and it always takes time to stop what you're doing, format the data in the manner expected by that particular data interface, and then resume your work.

With "remember", you just hit M-x remember (you'd probably want to bind this to an easily accessible keystroke, like C-x M-r), slam in your text however you like, and then hit C-c C-c. It will file the data away for later retrieval, and possibly indexing.

Indexing is to data what "studying" is in the real world. What you do when you study (or lucubrate, for some of us) is to realize certain relationships implicit in the data, so that you can make use of those relationships. Expressing that a certain quote you remembered was a religious quote, and that you want the ability to pull up all quotes of a religious nature, is what studying does. This is a more labor intensive task than the original remembering of the data, and it's typical in real life to set aside a special period of time for doing this work.

"Remember" works in the same way. When you enter data, either by
typing it into a buffer, or using the contents of the selected region, it will store that data -- unindexed, uninterpreted -- in a data pool. It will also try to remember as much context information as possible (any text properties that were set, where you copied it from, when, how, etc). Later, you can walk through your accumulated set of data (both organized, and unorganized) and easily begin moving things around, and making annotations that will express the full meaning of that data, as far as you know it.

Obviously this latter stage is more user-interface intensive, and it would be nice if "remember" could do it as elegantly as possible, rather than requiring a billion keystrokes to reorganize your hierarchy. Well, as the future arrives, hopefully experience and user feedback will help to make this as intuitive a tool as possible.

* Future Goals

This tool hopes to track (and by doing it with as little new code as possible):

 - The raw data that gets entered

 - The relationships between that data (either determined
   implicitly by parsing the input, or explicitly by the user's
   studying the data).

 - Revisioning of the data

 - Where it came from, and any context information that can be
   programmatically determined.

 - Allowing particular views of the initially amorphous data pool
   (ala the Xanadu concept).

 - Storage of the data in a manner most appropriate to that data,
   such as keeping address-book type information in BBDB, etc.

* Using "remember"

As a rough beginning, what I do is to keep my remember-data-file in outline-mode format, with a final entry called "* Raw data". Then, at intervals, I can move the data that gets appended there into other places. But certainly this should evolve into an intuitive mechanism for shuffling data off to its appropriate corner of the universe.

To map the primary remember function to the keystroke F8, do the following.

  (define-key global-map [f8] 'remember)

* Feedback

If Emacs could become a more intelligent data store, where brainstorming would focus on the IDEAS involved -- rather than the structuring and format of those ideas, or having to stop your current flow of work in order to record them -- it would map much more closely to how the mind (well, at least mine) works, and hence would eliminate that very manual-ness which computers from the very beginning have been championed as being able to reduce.

Have you ever noticed that having a laptop to write on doesn't
_actually_ increase the amount of quality material that you turn
out, in the long run? Perhaps its because the time we save
electronically in one way, we're losing electronically in another; the tool should never dominate one's focus. As the mystic Farīd ud-Dīn ʿAṭṭār wrote: "Be occupied as little as possible with things of the outer world but much with things of the inner world; then right action will overcome inaction."

* Diary integration

To use, add the following to your .emacs:

  ;; This should be before other entries that may return t
  (add-to-list 'remember-handler-functions 'remember-diary-extract-entries)

This module recognizes entries of the form (defined by remember-diary-regexp)

  DIARY: ....

and puts them in your ~/.diary (or remember-diary-file) together with an annotation. Dates in the form YYYY.MM.DD are converted to YYYY-MM-DD so that diary can understand them.

For example:

  DIARY: 2003.08.12 Sacha's birthday

is stored as

  2003.08.12 Sacha's birthday

Defined variables (33)

Defined functions (18)

Defined faces (0)