Re: One vs many directories

[Jean Louis <bugs@gnu.support>]

* Tim Cross <theophilusx@gmail.com> [2020-11-21 13:15]:
> I have used similar tools in the past. However, what I find frustrating
> about them is that your now dependent on another bit of technology - a
> database of some type with all the issues that adds - installation,
> upgrades, maintenance, backups etc. The thing I like best about Org is
> that in the end, it is just a collection of plain text documents.

Thank you, I understand your opinion. Let me share my viewpoint.

Installing database such as PostgreSQL and configuring it is way more
simpler than installing Emacs and configuring Org mode or Emacs
itself.

It goes like this:

$ PACKAGE-MANAGER install DATABASE

Later you just do:

$ createdb MY-DB

and then insert tables, usually program does that for you.

- installation is simpler than Emacs, definitely, unless you install
  from sources which is in turn about the same as Emacs

- upgrade is not necessary in general, majority of users need not ever
  upgrade. GNU/Linux distributions make upgrades easy and flowless
  normally.

- for maintenance of database I do not know what this refers to. To me
  it means working with the database. Inserting, updating, editing
  records. That is writing text. I write text in Org and I write text
  in the database, it is about same thing without file on file
  system. It lacks maybe versioning system which in turn can be
  implemented easily to simple first backup the field into versioning
  table before editing such. Simple.

- backup of the database is way easier than backup of the file
  system.

export POSTGRESQL_BACKUPFILE="/home/data1/protected/Security/Backup/Database/PostgreSQL/`/bin/date -Iminutes`.sql.gz"

alias backupPgDB='nice -n 19 /usr/local/bin/pg_dumpall | gzip > $POSTGRESQL_BACKUPFILE'

That is all what is needed for the backup in my example. It can run by
cron job that I do not even think about it. I can send database file
by email or encrypt it by GnuPG and upload to some remote server
automatically.

> I also wanted to mention another Emacs project which is quite
> interesting and has actually been around a lot longer than org,
> Hyperbole. I've not got a URL handy, but I'm sure you can find it with
> google.

GNU Hyperbole
https://www.gnu.org/s/hyperbole

I am using Hyperbole everyday and it integrates with Org mode and can
use Org links. It is on meta-level in comparison to hyperlinking in
Org mode and in itself it is hypertextual information management
system but is only in one part comparable to Org. It is not a mode for
editing files, but that one part is named Koutliner and that is
outline mode comparable to Org with one major difference that
Koutliner provides specific ID for each part of the nodes in the
substree.

People often mix GNU Hyperbole and Org, but they are not comparable as
Hyperbole is Hyper and Org is Org. Being Hyper it works in every file
or every directory. I just wish to find out how to create buttons on
the fly to interact between Hyperscope and Hyperbole. 

> It is an interesting system which pretty much makes everything
> possible in a document a hyperlink. Provides some very interesting
> ways of linking documents.

Yes. When I enter my project directory there I keep ./HYPB directory
button file. Then there are some tasks which are easily invoked by
Hyperbole that relate to various maintenances:

General actions
===============

- <(Record the Developed Traffic incidents)>
- <(Create General Log entry)>

Search
======

Database actions
================

- <(CF: Last 200 contacts)>
- <(CF: People with mining lands)>
- <(CF: Create contact and edit it)>

Database settings
=================

- <(Turn off speech)>

Database Maintenance
====================

- <(CF: Find largest contact names)>
- <(CF: Find possible doctors)>
- <(CF: Normalize public email addresses)>
- <(CF: Normalize phone numbers)>

Now I am sure that same can be done with Org file as it offers many
kinds of hyperlinking. While Org hyperlinks can be placed in any file
and used in any mode they look ugly in non-Org text files without Org
links mode enabled.

> My preference has always been to 'do my own thing'.

My preference for other people is same, that each should find their
way and use the tools and paradigms they find familiar or appropriate.

> I tend to look at other information management approaches and cherry
> pick the bits which I like and then replicate them in org. I don't
> find org as limiting as others seem to, but I'm also quite happy to
> add in my own elisp to tweak it the way I want it to be - thats why
> I love emacs.

Of course. Me too. And I develop my own. I like to integrate
things. For example tasks in Org mode I have to delegate to people so
I click a button, choose person, choose valid email if there are many
and task is sent to person. Person gives me feedback and I update the
task conducted by that person. That is integrating.

Maybe I am in the Org file for the person, how do I call quickly this
person? If there is #+MACRO: contact-id 239917 then I can just click a
key and send SMS or initiate a call straight from the object related
to that contact in this case Org file. And it should work for any
object any file related to person or other object.

If I inspect country database, maybe contacts are related to country,
from country I should be able to find contacts of that country. From
any file or location in directories related to person ABC, I should be
able to initiate communication with such user without again finding
user's n

I am in process of transition to HyperScope as dynamic knowledge
repository.

As Org files are too many. Some are important, some not, some carry
this and that information, it is everywhere, not any more simple.

I am sorting it automatically into people's directories. Contracts are
written by Org and such contracts are related to people, they shall be
sorted by relations.

This applies in general to any kind of files. When files are sorted
they should be in the same time, same second, also indexed. Automated
indexing should also help the user. I do not mean indexed as full
file, just as meta information or hyperlink to which one can find a
quick reference.