Re: Remembrance Agents

[Jean Louis <bugs@gnu.support>]

* Gerardo Moro <gerardomoro37@gmail.com> [2020-11-30 09:49]:
> Thanks! I get an idea. Will try to use soon and let you know.
> Thanks for the links, very helpful.
> Question: when you say it uses emails, can it read any email database (I
> have old ones from Microsoft outlook both Mac and Windows)? What if you
> only use webmail, in which format shall you feed the RA?

I think that all those formats can be converted to something you could
be using centrally on your computer.

I am recommending Maildir format. But that would depend of your email
client that you use.

I am using Maildir since years as each message is its own file and
concurrency is there and I never lose messages. I was using mbox and
mh and various email formats, nothing was ever so stable as Maildir.

Putting each person's conversation in separate folder:

~/Maildir/name@example.com

then allows me to easily access all previous conversation to that
person. This implies that saving messages must be configured to save
into such specific maildir automatically.

recoll is not remembrance agent. It is a desktop search tool that
supports various formats. There are other desktop search tools on
GNU/Linux as free software:

Beagle - Quickly find the stuff you care about.
http://beagle-project.org/

Recoll is a desktop full-text search tool.
https://www.lesbonscomptes.com/recoll/

Tracker is a filesystem indexer, metadata storage system and search tool
https://wiki.gnome.org/Projects/Tracker

NEPOMUK - The Social Semantic Desktop
https://nepomuk.semanticdesktop.org/

Terrier is a highly flexible, efficient, and effective open source search engine
http://www.terrier.org/

In general as Emacs user you may need to choose one that works on
command line. I am not sure if all of those work on command line.

Command line output can then be customized to appear in Emacs.

Then hyperlink may be constructed in Org file as well for various
relevant semantic searches.

For Hyperscope dynamic knowledge repository since we talked about
remembrance I have at least implemented the table column hlinks_rank
that increases its rank by usage. Maybe I should call it different.

Org mode does not track every specific node and is not multi-user:

- when was the specific node edited?
- which person edited specific node?
- what is previous version of that node?
- which hyperlink was accessed how many times?

Would it track that information without disturbance, then users would
have list of most used nodes, most accessed or frequently edited
nodes, groupware based nodes.

In that specific sense of tracking which URL have been accessed how
many times, this gives a list that helps to remember what could be
more relevant in context user is working in.

As the rank increases, the hyperdocument has its own tags, title,
information, and those words mostly used can be indexed and while user
is typing those words could appear on the side to give more popular
references.

Then access could be tracked on files as PDF, images, videos,
directories within Dired, then last of most accessed files could be in
front of us. Recentf does similar function in Emacs. But Emacs
bookmarks are not tracked and sorted by their access.

I will try downloading Remembrance from Debian, as maybe they have
patches that work.

Jean