Re: One vs many directories

[Jean Louis <bugs@gnu.support>]

* Ihor Radchenko <yantar92@gmail.com> [2020-11-24 10:57]:
> > I find it entertaining for now. Now, what is exomind? 
> 
> Unless I misunderstood, Jean referred to "external brain" concept:
> - https://beepb00p.xyz/exobrain/

The more you send me reference more I discover other set of people
doing same what I am doing. Since I have implemented central meta
level organization it is moving rapidly, everthing gets sorted. It
develops by itself and is rapidly accessible.

That website I have to mirror locally to pick ideas and learn from
others. Mirroring I do with:

$ wget -Emk http://example.com

As that command replaces all hyperlinks to local hyperlinks. That
person advanced in organization of things. I stick to few principles
and just design it by principles.

Design works rapidly. Few Emacs Lisp functions and access to reports
listed in Emacs Buffers and integration with other tools.

With one function and one PostgreSQL table defined in 3 minutes I get
rudimentary backup and version system for any column values that I am
editing in the database. If I edit note, the note is versioned
(previous version stored) before I start editing it. Principles I am
following are basics what programmers like, to minimize or eliminate
repetitions and efforts to achieve the goal.

Person above have extracted or exported its own database of hyperlinks
to hyperdocuments. My side I have made for now Org export of any
subtree or the whole dynamic knowledge repository. There are many
things to go. In Emacs development version all kinds of hyperlinks can
get their handlers like gopher:// gemini:// message: tel: sms: and
htat will be very helpful.

No, I do not use "exobrain" as a term. I rather lean on Engelbart's
terminology and follow his principles as we are very late to implement
what was envisioned back in 1968 and before. It is 52 years already.

And many more years since Memex has been invented:

Memex
https://en.wikipedia.org/wiki/Memex

As author said: "The memex device as described by Bush "would use
microfilm storage, dry photography, and analog computing to give
postwar scholars access to a huge, indexed repository of knowledge any
section of which could be called up with a few keystrokes."

And that is exactly what I am creating here to have anything called up
with few keystrokes and to be able to share files with individuals or
groups of people without more thinking but just designated what to be
done.

Have group of 5 people to share notes with? Just find the designated
group and click share. Computer would handle the rest, maybe send
files by emails individually, maybe inform people by SMS, maybe upload
files and share password protected hyperlinks with those people.

Integration is another keyword I like to follow. Android principle of
sharing is pretty much based on integration. We have all the small
functions around us only not well integrated with their relations that
concern human problems.

We have files on file system which we cannot easily share with groups
or people we want. Address books are all sparse, one is in this email
client, one is separate, one is on the mobile device, another email
client does not synchronize, and so on. I have forgotten this long
ago and use central address book from where everything derive:

- no Google, clouds, etc. that is very insecure. Do not give contacts
  to Google, there are hundreds of thousands of staff members there
  and no guarantee whatsoever that they will not read it.

- keeping contacts on my computers. I have already spent money for
  hard disk, there is enough space

- exporting contacts from central database and importing to email
  clients, mobile devices, this way everything is synchronized.

How quickly can GNU/Linux user share a file with somebody?

- locate the file by using hierarchical browsing. If file system is a
  mess, this alone may take some time

- open up email reader

- find that email address. If it is in the email reader already it is
  good. But it could be in the phone. It could be on paper, or on
  business card. Where is it? Maybe calling person? But where is the
  phone number? On first phone, second phone... if all is synchronized
  maybe is easier to find.

- attach the file

- send the file.

But then sending SMS or calling in the same time does not
work. The above process is not well integrated.

It could work like this:

- user just thinks of what has to be shared with other person, types
  the terms related to the thought

- locates the file and press share

- locates the user and press enter. FINISHED

That would be better integration. Even better it would be if user can
choose the automated workflow option:

1. send the file, automatically record that file has been sent to
   specific user. Tell user automatically how many files are attached
   and attach annotation belonging to the file as body of the email or
   any instructions.

2. in the same time inform the user by SMS that file has been sent and
   record that SMS have been sent. Software like kdeconnect, gnokii
   can be used for it.

3. within 1 hour, or other period of time, computer asks to initiate
   the call to the user to follow up about the file sent and maybe
   nudges few times and records the action. Software like termux tools
   can be used for it.

My first big surprise with Org was that there was no possibility to
assign the task to other person and send that task! I actually could
not believe that it was meant for single person or personal tasks and
notes. Then I made the function to share the task quickly to any
person assigned to the task. If person is assigned, task is sent to
the person. If no person is assigned then I choose to which person to
send it. This includes also groups of people.

> - https://zettelkasten.de/posts/extend-your-mind-and-memory-with-a-zettelkasten/

That is similar idea of organizing. There is claim that one shall
forget about categories and rather use tags. I think that using any
types of attributes is better and using more attributes helps in
quicker location.

> - https://github.com/novoid/Memacs

I have installed it and not yet used it. I would not like having too
many tools on file system to manage information. There are too many
memacs tools made for console. In general I am tracking all SMS sent
from phones to other people, they are automatically inserted into
corresponding people's objects. Then I know which people received
what. Phone calls can be tracked too. Phone calls can be
recorded. Sales and marketing departments need that. I am using now
only principles from Memacs and implement some of them in Emacs
Lisp. As I like integration I do not like external tools, but the
dynanic knowledge repository must be usable externally without
Emacs.

That becomes very easy by expanding the whole tree of notes into the
file system from time to time and generating meanings for symlinks
that point to fixed locations on file system.

The centralized subtrees or nodes of my dynamic knowledge repository
can be moved easily from one parent node to other parent node. This is
because human must sort things properly. But if such are pointing to
files on file system those files never change. Meaning and relations
can change but file location should not change. Directories are more
static then files. They would never change. Files if not indexed but
located in the archive could maybe change or get updated.

Git repository could get updated but its directory need never change
in the future.

Let us say there are many PDFs to be indexed and accessed through
semantics. The PDF file name could change but access to PDF file need
never change. Renaming PDF file need not change access to PDF file in
other word there is no need to rename it twice, it can remain in the
database and get accessed automatically. But that requires directory
to be static, or it requires md5sum of the file to be
static. Something must be static that file can be found by the
system. Best is when file is under specific unique ID that never
changes. Then everything becomes unique and clear. And symlink can be
automatically generated:

~/hyperscope/1/2/3/432.pdf would be file

~/hyperscope/1/2/3/Knowledge.pdf would be symlink to 432.pdf
automatically generated and from time to time updated if there were
many changes in the database.

The Org hyperlink to the file could point to:
~/hyperscope/1/2/3/432.pdf because file location is this way static
and will never change.

But the Org hyperlink could as well point to meta level hyperlink
(hyperscope 432) as that would open the file no matter where is the
location.

And if file is on remote server, something like

[[PDF File][(hyperscope 432 2)]]

would then work quite well. As this type of dynamic knowledge
repository is multi user automatically as database is multi user. It
means files can be accessed from all over the world and groupware
collaboration becomes trivial as PostgreSQL is networked database.

Now for the user accessing the specific database then the hyperlink
can be just (hyperscope 432) and user who is remote could say by
activating the hyperlink that remote user likes to have the
file.

Program must know if user is local or remote. If user is remote the
file can be sent by email, it could be automatically encrypted and
sent, encrypted and uploaded to web server, or uploaded to web server
with password encrypted access or without. Any information can be
protected and not all information need to be shown on public
webservers.

Maybe it becomes better to use the URI like:
hyperscope://user@example.com:432

whereby 432 would not indicate the database port but rather the ID of
the hyperdocument to be activated or accessed. user@example.com would
have relation to the actual username, password, hostname, database
name and port on the user's own system and program installed as such
without local database. Maybe port could be optional as multiple ports
could be on the same hostname.

That would create unique access to specific domain and specific user
on remote hyperscope server. It becomes possible to securely share and
access files or do any action on such files by using the groupware
features.

The big difference with the WWW is that system is structured and
offers liberty on how to access files and what to do with the system.

User may remotely invoke emails to be sent to groups of people. This
is not what WWW offers by default.

User could remotely edit Org file only by using the database. Org file
need not be located on the file system. And yet such Org file can be
automatically saved on the file system or sent to other people.

Locating 4-5 or more people becomes possible, marking of them and
quick export to Org file becomes possible. Then user may invoke
further actions such as visiting people, negotiating, calling people,
sending them information by post, sending SMS to people.

When backed up by well networked database it becomes multi user
collaborative meta level Org system. 

> - https://blog.jethro.dev/posts/org_mode_workflow_preview/

I have captured that one for later research. Small details and notes
do matter when creating some new useful features.

Jean