knowledge

A knowledge daemon to collect ideas and auto organize them

View project on GitHub

About Knwoledge (STILL DRAFT)

The muse visits during the act of creation, not before

Knwoledge Goals

Needs and past experience

Primary need is to collect and auto organize (at least “facet”) news I read on Internet.

My primary source is Hacker News but I collect data from other souces too. I do not use bookmarks: I tend to publish interesting articles on twitter, or sent it to a special “todo” mailbox for further reading. I publish only things I think worth to, and I send to my todo-inbox only things I plan to use on my projects. I use “Read Later” iPhone functionality when I have a lot of mid-priority stuff to read. Eventually when an idea is ready for publishing, I write an article on my Blog.

Some Ideas can be private, but as a primary goal should be the abiity to prepare a “Idea ReCap” with all the information is presented to be able to write an article on it.

  1. Search is a crucial need.
  2. Aggregation from diferent sources seems unaivodable
  3. A lot of “bookmark” application tried in the past, they seems the wrong direction
  4. A long browser history seems a good source for the article basic link collection

Past experiments

A Wordpress site with the folloing plugin offered a good starting point:

  • dw-question-answer to create a StackOverflow-like tassonomy
  • elasticpress to index site via elastic search
  • wpematico to aggressivly collect rss information (can be a bit dangerous for intellectual property)
  • auto-upload-images to collect external images, schematics
  • buddypress /buddypress-docs for optional social engage
  • Simple Twenty Twenty theme with a fixed page showing:
    • Calendar posts
    • Categories

Other fetaure included:

  • Able to import from other wordpress site
  • Stable XML-RPC API to quickly enable thrid parts post processing

Candidates

  • Wordpress +Elastic Search + Jupiter
    • Pro
      • Nice look
      • Resouce consumption medium (1.2GB RAM)
      • XML API
    • Cons
      • Plugin support lacks
      • Heavy MySQL DB
      • Attack surface not small
  • GitLab
    • Pro
      • Can set up an Editorial boards based on Issues
      • Can Manage wikis via simple github-rendered projects
      • Code-centric
    • Cons
      • Memory consumption heavy (4GB RAM)
      • Less nice themes
      • Heavy Database
  • Custom Python aggregator
    • Pro
      • Very small resouce consumption
      • Portable sqlite3 database