OrgNV - alternative to deft

OrgNV - alternative to deft

[Juan José García-Ripoll]

Hi,

I have started a personal project, with no aims to create a package yet,
to create an alternative to deft. I have very specific reasons:

- Open the library to being used programatically, similar to how
  completing-read works.
- Built around 'grep', to allow more sophisticated queries in the
  future.
- No databases. No "Xapian"-like interfaces. No cache files.
- Multiple source directories.
- Focus on org-mode nodes.
- Easier interface to org-mode for creating links directly from an
  org-mode file.

The result feels pretty snappy. Reading 1000 files with 'grep' and
parsing them takes less than 0.5s in an old laptop. It is so fast that
I don't really feel like adding a cache around it.

I feel this paradigm can scale to more interesting features, such as
querying #+KEYWORD or tags in files, which I am using in my personal
knowledge database.

Feedback is very much welcome. This is the first such "interactive"
project I start with Emacs and I am unfamiliar with the programming
conventions.
http://juanjose.garciaripoll.com/blog/orgnv-navigate-notes-with-grep/index.html

Cheers

-- 
Juan José García Ripoll

Quantum Information and Foundations Group
http://quinfog.hbar.es - http://juanjose.garciaripoll.com

Re: OrgNV - alternative to deft

[Bob Newell]

Juan,

You might take a look at my 'scraps' package. I created it a
little while back and it does a fair amount of what you
mention.

I've got over 3,000 files taking up over 64 MB and it's still
quite fast.

http://www.bobnewell.net/filez/scraps.zip

-- 
Bob Newell
Honolulu, Hawai`i
- Via Gnus/BBDB/Org/Emacs/Linux

Re: OrgNV - alternative to deft

[Juan José García-Ripoll]

Bob Newell <bobnewell@bobnewell.net> writes:
> You might take a look at my 'scraps' package. I created it a
> little while back and it does a fair amount of what you
> mention.

Thanks, looks very interesting. Does some of what I need, but I differ
in some design choices:

- I prefer the incremental search scheme a-la-deft instead of dired.
- I also do not care for file names and prefer to have editable titles /
summaries.
- A reason for this is that I care about linking and backlinking
notes. Fixing file name <-> title association makes the database static
and difficult to edit.
- I am also into having multiple folders for separating big categories
and areas of interest (hence the recursive approach).
- I also would like to attach multimedia / PDF's / images to the notes as
linked or embedded assets. This is easy if directories are allowed and
files filtered by extension.

In any case...

> I've got over 3,000 files taking up over 64 MB and it's still quite
> fast.

... your work proves that using grep is not such a bad idea after all
because of available memory and SSD's, as you say in the PDF.

-- 
Juan José García Ripoll

Quantum Information and Foundations Group
http://quinfog.hbar.es - http://juanjose.garciaripoll.com