2.0 KiB
How I Organize My Research Archive
Introduction
As (acting as of 04/14/2021) CTO of an incredibly complex product/services portfolio, I have a large amount of research material. I am writing this note on how I organize it and make it fully available to the world without any overhead on my part.
Short version:
Documents
This means PDF, print to PDF of web page, other file types (doc/xls/whatever)
- Bookmark the link where'er is appropriate in my taxonomy.
- Download the URI to ~/ResearchMaterial
- calibre (on Raberry pi) is using ~/ResearchMaterial and that folder is synced via nextcloud to the corporate file server so I can also use Docear/Polar/Zotero on my x86 vm
NOtes
I use this git repository (notes-public) with a folder of markdown files.
I primarily use VsCode on the ras pi todo note taking (and of course longer document creation) Occasionally on mobile (iOS), I use Buffer and the WorkingCopy git client to make quick edits or capture something when I"m out and get inspired.
I heavily use Working Copy on my iPad Mini 5th Generation to do code review, issue cleanup etc and I use Calibre to load up a long research paepr and read it over. I also may access the folder directly via samba if I want to annotate.
Source material
More detailed version
Taxonomy
I use a single taxonomy across:
- e-mail folders
- bookmarks
- home directory
- notes folder (this repository)
It is:
❯ ls
Board CEO CFO CIO CMO COO CTO dotfiles dotfiles-git landed notes-public personal PFVEMer
~/charles 15:03:47
❯
- Board
- CEO
- CFO
- CIO
- COO
- CMO
- CTO
I then have division names or project/product names under those folders.
More coming soon , still dialing things in.