Re: [OT] The Brain

[John Hendy <jw.hendy@gmail.com>]

On Sat, Apr 23, 2011 at 12:39 PM, Marcelo de Moraes Serpa
<celoserpa@gmail.com> wrote:
> Hey list,
>
> Have you seen this?
>
> http://www.thebrain.com/

Yes!

>
> It's a mind-mapping-pim-everything-in-one kind of visualization
> software. It seems pretty interesting - I like the way you can connect
> the items at several levels, helping to build rich relationship
> information.
>
> I've been thinking on building something like this for org, where
> links in the files would provide the relationship information and we
> could then export as a freemind map (or any other format that easily
> allows describing relationships), for example. The current exporter
> does this, but only for individual files, the idea is to recursively
> build it for a specific org directory (where multiple .org files are
> found)

*Note:* this got long, but it's because it's an exciting topic and I
was excited to share my thoughts. Hopefully it's entertaining or
helpful or at the very least not burdensome to your inbox...

Even without having an exportability, I have often longed for some
kind of "semantic hierarchy" for tracking information. That's the cool
thing about how Personal Brain works -- it links thinks to as many
places as you want vs. a traditional linear hierarchy. I think
something like this should be able to be accomplished by use of tags.
I actually tried a semantic filing system structure a bit back where
my plan was to name my files like so:[1]

proj_name-of-the-file_yyyy-mm-dd_[tag1-tag2-tag3].ext

I created a bunch of test files in this format and then used a shell
script to create a hierarchy of symlinked folders so that every file
was in every folder that applied. For a simple case, a structure might
exist like so:

Files:
- proj1_file1_[tag1-tag2]
- proj2_file2_[tag2-tag3]

Resulting hierarchy:
- proj1
--- tag1
------tag2
--- tag2
------ tag1

- tag1
--- proj1
------ tag2
--- tag2
------ proj1

- tag2
--- proj1
------ tag1
--- proj2
------ tag3
--- tag1
------ proj1

- tag3
--- proj2
------ tag2
--- tag2
------ proj2

That's the directory hierarchy only -- not listing the files. Each
file will appear in *every* folder that matches it's project or tag (I
was even going to go by extension as well and add in dirs for docs,
pdf, etc.). So, a little crazy, but the test run was actually really
neat. You can get to any file via any number of paths.

This was my idea for work, because I often organize things by project,
but get annoyed when I can't find what I'm looking for because it's
elsewhere. For example, my company has an invention submission system.
You do something in the lab, think it's worth attempting to patent, do
background searching, and then write up an Invention Submission. That
gets reviewed by a committee, and if they decide it's promising, you
present to a manager who approves the filing fees.

I have filed several of these submissions and keep the supporting
pictures/videos/etc. in the folder with the submissions, and those are
in a directory called "ip." One day I was making a presentation and
hunting for a video I *knew* existed but couldn't find anywhere.
Whattya know -- it *was* for that project, but it was in the ip
folder. Bummer. I eventually found it, that's what gave me the idea
above.

proj would be short 3-4 letter abbreviations for projects.
tags would be things like ip, vid, reuse, pic, draw and would help tag
various uses for the files. If I tagged things, I could look in the
video symlinked directory and find what I wanted. Or go to videos ->
project name and that would narrow it down even further.

Now, I wasn't actually going to move *anything* -- just create an
independent folder structure of symlinks created and updated by
running a bash script that would find all unique combinations and then
create dir structures to match.

To the point (and thanks if you made it this far): I would *love* for
something like this to be available in org, and it would give things
the feel of The Brain. Imagine being able to document your information
in typical headline fashion, but enter some separate viewing mode by
tag to find what you're looking for. Things just *aren't* linear,
often times. They double count in two or more "meta" categories --
projects, type of information, a todo, reminder, the "class" of
information (ip, personal note-to-self, something to read, report),
even how the information came to you (meeting minute?, website, etc.).
I think it would be slick to enter a "brain" like viewing system that
used tags or properties to swim through files in your agenda. I'm
often having to recall some tidbit of info and rack my brain for what
context I wrote it down in if the agenda search isn't working.
Eventually I'll remember what the meeting was about where it was said
and find it.

Plain and simple, what I write when something happens isn't how I'll
necessarily recall it later, hence search phrases not always being
useful. I think "bucket" tags (like the "bucket" nodes in Personal
Brain) would be very neat indeed.


[1] There are some systems that kind of do this already, but most of
them wanted to "suck" my information into their own database and I
really, really didn't like that. Here was my attempt at getting help
:) http://stackoverflow.com/questions/4081230/bash-for-semantic-file-structure-creation


Best regards,
John


>
> Food for thought :)
>
> Cheers,
>
> Marcelo.
>
>