Ask HN: What is your knowledge toolchain in 2018? | Hacker News Reader

Ask HN: What is your knowledge toolchain in 2018?

10 pointswcrichton7 comments

I've found it increasingly difficult to engage with and keep track of the sum of knowledge I encounter every day between papers, talks, podcasts, blog posts, and my own thoughts. Specifically, I care about the following tasks:

1. Active engagement: while consuming a piece of knowledge, I want to write down my thoughts/reactions with as little friction as possible. For visual knowledge (web pages and PDFs), I want to annotate my thoughts inline (preferably with support for stylus input).

2. Summarization: after finishing a piece, I want to review my annotations and pick out the important ones, as well as potentially write a standalone summary of the piece. I also want to categorize the piece via tags.

3. Linking: I want to be able to provide different kinds of links between knowledge pieces, either implicitly (e.g. by providing the same tag) or explicitly (e.g. "this web page is the HN comment section for this article"). When writing comments, I should be able to write down internal hyperlinks to any knowledge piece in my collection.

4. Contextualization: I want to compare/contrast the main points of the article with linked articles. I should be able to easily search along every dimension of my data cube (e.g. all articles from source X, all articles about topic Y, etc.) to produce different contexts.

For lack of a true memex, my question is: what tools do you use today to accomplish these tasks? From my experience, almost everyone has a different ad-hoc setup, so I'm curious to understand the diversity in HN's community. I've seen people use tools like Evernote, Scrivener, or even a custom MediaWiki to perform a subset of these tasks, but I've never seen a complete toolchain.

I'm also interested in perspectives on: should this problem be approached in Unix fashion (each task deserves a single tool that does it well?) or in Emacs fashion (the entire toolchain needs a single, over-arching, opinionated framework for maximum integration)?

I once pursued a memex solution like you describe and found that a simple static website generator worked well, but was high friction. You had to format notes, move files around, link images, etc.

My experience was that the collection of knowledge didn't produce many valuable insights or actionable results on it's own. I often found myself hoarding information, and creating a large backlog of things to process.

Summarization, linking, and contextualization are valuable, but then you have to do something to make that effort worthwhile, something that adds value to your life.

The more difficult part of this toolchain for me, was actually slowing down and thinking deeply about problems and connections. This required a disciplined habit of writing in a distraction-free environment (offline).

Once I realized the value of slowing things down, developing habits of reading (paper) books, and spending several hours a week writing and thinking, this diminished the value of having a fully featured knowledge tool chain for me.

I keep text files in a simple directory structure, and use cmd+t and project search in Sublime Text to find things.

That said, I'm still interested in a solution to this problem, since I spent no small amount of time is search of the grail.

The last 6 months or so I've been using TeXShop to make LaTeX files of everything I'm learning about, i.e. to produce beautiful-looking pdfs. It's going great. I have 2 or 3 files I'm working on currently; there's another few on specific themes, which will be books or papers. One is a daily diary-type one to put everything I'm learning in: names of books, papers, videos, extracts, pictures, links, quotes, extracts, my thoughts. Chapters in that one spin off into their own file when they get big enough. Another is on all the little LaTeX/TeX things I've been learning about.

Before I started doing this, most of the daily stuff I do - papers and books read, thoughts about it, good pages/paragraphs - would be lost. And I have a bunch of old notebooks with everything disorganized that I've never looked at again. I'm into so many different things, keeping records about it has always been a problem for me until now.

I know, sounds weird huh. Things I like about this system: hyperlinks, internal links, footnotes, endnotes are super-easy. Anything that isn't super-easy, you can just make a new LaTeX command or environment to make it super-easy. Easy to insert pictures, cropped or scaled. Mostly, the compiled pdfs look so nice, it's a great pleasure to work on them. TikZ is sometimes the best way I know to make professional-looking diagrams of stuff. So yeah, learning LaTeX (and basic TeX) isn't much fun, a few months of hard work, but StackExchange and a few LaTeX books makes it fairly painless. There are included packages that do most things you could want, the documentation that comes with them is almost always impressive. (Compared with man pages, at least!) Most of the learning time is learning how to use all these packages. I'm no expert, but I've stopped consulting SE and books; I know enough for my purposes. Also, the stuff ends up in in a virtually ready-to-publish state.

Also I write quick plans, diagrams, pseudocode, ideas etc with ballpoint pens in notebooks and on graph paper.

Mac user. I use OneNote for 1) active reading. I use it to import articles so I can annotate them. I use skim for PDF files.

For linking & collecting I use DevonThink.

Summarization markdown files using iAWriter. It can handle adding pictures very well. I make heavy use of screenshots.

To collect ideas I use Scapple, it's a fantastic brainstorming tool.

I use Monosnap to take screenshots. I can drag and drop and picture into either tool I mentioned above.

The most important step, however, is the last one. Putting my own thoughts down into a final text (that I hopefully publish somewhere)

There is no single tool. You could use DevonThink for everything I guess.

Having said all this, be careful not to get hung up too much on the tools, they can be a mega distraction (productivity app p*rn). Use what works for your flow and try to produce your own thoughts, which is hard.

(Apple user, so mentioning apps for this ecosystem)

While listening to podcasts or get an idea while walking around, I record using the iPhone Voice Memos.

I then reprocess those recordings from time to time and put the gist into written words.

When I'm doing some brain dump on random ideas I use MindNode Pro.

Notion (notion.so) is a pretty great app for keeping a record of everything, from random text to lists and tables. Has cool interlinking features and works on Web/Mac/iOS (and possibly others but I didn't check)

I try to things as simple as possible, and do all my note taking in Google Docs. Sometimes I worry that I depend too much on Google, so I started to take notes in plaintext using Vim and uploading them to my GitHub repo. For everything else (including visualization) I use pen and paper (or a graph composition notebook).

Not exactly what you're talking about, but the most important tool I use for ingesting/remembering things is Anki: https://apps.ankiweb.net/

I have started doing something along these lines because I have the habit of dabbling around with new shiny programming languages and libraries. I also consume a lot of information online -- blogs, articles, and ebooks of things that I am interested in.

Previous to this, I had a paper based notebook to hold all my thoughts and whatever I learned. The problems that I faced with that setup were:

- For offline notes, I lose out the luxury of editing cleanly and indexing. Writing technical notes on paper go out of hand after some time. (I still have a paper based journal and personally find paper to be better in all other respects but I digress).

- Because indexing is tough, so is searching. I can make a hundred tags, but if I lose out on giving that page an entry in index then it becomes difficult to find.

I have been an Emacs user for some years now and have gradually moved my setup to org-mode[0]. I arrange my thoughts/knowledge across various files and folders, give tags to the data and searching and arranging this has been pretty solid. Searching is just a single command "Ctrl-c p s g" and it greps through all the entries, e.g: https://imgur.com/a/IY0TE3Y

Coming to the points that you mentioned, I do the following

1 & 2. Active Engagement & Summary: With a dual monitor setup, Emacs is (almost) always opened. If I come across something, then I quickly open/create the associated file and log my thoughts. This is true when I'm at a computer, both at office and at home. Otherwise, I try to make an entry in some note taking app on my phone (I've been using ColorNote for Android for last 6 years or so, it syncs and backs-up the data). This phone based approach is something I still am not 100% comfortable with, but at least I have some place to dump my ideas.

3. Linking is done by org mode tags[1], and the file/folder structure that I have on disk. For metadata (webpage, context, etc.), we can associated a property drawer with every entry.[2]

4. Contextualization: Maybe a combination of tagging and searching can do that? I'm not an expert with org, but it supports searching tags with boolean operators[3], so that might float your boat.

This setup is plain text, and supports encryption via Emacs (and auto backup if you save in Dropbox). I have a private repo at gitlab and push every couple days.

[0] https://orgmode.org/

[1] https://orgmode.org/manual/Tags.html

[2] https://orgmode.org/manual/Drawers.html

[3] https://orgmode.org/manual/Matching-tags-and-properties.html...

Reply on news.ycombinator.com