How do I use PKMS properly?

How do I become productive? #

Productivity is one of those things where people, especially someone like me, wants to attain, but is so vague and general that it’s hard to pinpoint how to become good and effective at it. Yea, I want to be productive because I have certain goals in life, and you can’t achieve those without some work. But how exactly to be productive? I promised in May that I would make a monthly review of chemistry papers. Turns out, my current setup makes it hard to write up a new review article. In fact, I don’t know what should it look like.

Currently, I’m using Obsidian as my personal knowledge management system (PKMS), and at work we just use a measly excel sheet to track personal stuff. Kinda ew, but it works. I should try learning how VLOOKUP works.

I digress. Going back to Obsidian: I currently use it to write my blog posts, save some web clippings, annotate some pdfs via Zotero, and make some notes out of those. Using this method, I just felt that it was very tedious. Two apps with poor integration does not really work well for me. However, storing pdfs inside Obsidian seems very iffy. I don’t know what exactly should I do here.

After annotating my papers, I export them to obsidian, which just gives me a single note full of all the annotations that are somewhat formatted poorly. I guess I was expecting to be better formatted with each annotation in its own note. That’s probably too much? But I like the idea of granular/atomic notes.

Uh, what was I going for again. Ah right, becoming productive.

Essentially, the problem I can identify right now is that my mind is a mess and I am prone to reading articles without absorbing and remembering. Like reading for reading’s sake. My goal is to retain these information so I can actually remember and go back to these articles that I read beforehand.

So. What I need is a system. This system should help me manage all the information I come across with so I can go back and refer to these. Maybe a personal wiki of sorts. Some sort of knowledge database.

What are PKMS (to me!) #

Oops yea I just mentioned personal knowledge management systems candidly lol.

So I don’t want to rewrite what’s basically in the Wikipedia article, so let’s just define PKMS in my terms.

My goals are as follows:

1. To organize new information properly.
2. To retain information.
3. To use the information for writing blog posts (and other stuff).

These goals seem nice, but it doesn’t really describe anything what I want in a PKMS. For example, a binder notebook would solve these goals, but my brain isn’t compatible with just writing stuff on a notebook. Here are my gripes with a notebook:

First, information is written in a chronologically linear way. It’s hard to make corrections, edits, and updates without scribbling stuff or scrapping pages altogether. I personally believe that PKMS should be easy to update and doesn’t require a linear way of writing and reading information.

Second, notebooks (and honestly, most wiki-type/PKM software) are mostly built as a reference-like source of information. What I mean is that, yes, it is best when I need to refer back to when you need to remember a specific topic on demand, but I also like it when I commit important knowledge into my memory. Perhaps one with good integration with Leitner boxes/Anki would be nice? Back when I was studying for my board exam, I wrote all important topics in a 3”×5” index card. Is there a system that can consolidate information like that in a way that doesn’t get outdated too quickly?

Third, extracting raw information and elucidating it into proper literature is hard. I have a lot of blog ideas but they all get blocked by me doing inadequate research. How exactly should I approach this? Ideally a PKM system would help clarify the raw data I collect. My current approach of extracting info directly via Obsidian Web Clippings and Zotero seems so clunky. If I use the Zettelkasten method of organizing data, I find the step between reference notes and synthesizing from them very clunky. Is collecting random data then recognizing patterns from said data an easier task over something like having a goal of collecting specific data based on a pattern I think exists?

For example, if I want to research the topic of electrochemistry because I was interested in the topic of catalysis of certain reactions via electricity, then where do I start? Do I search up “electrochemistry” on Semantic Scholar or whatever and then save a bunch of papers to read later? Not really motivating to me. How about when I want to keep up with the current papers being published everyday? I think I can understand the abstract, but then truly understanding these things require some base foundational knowledge established beforehand, which I don’t have. I would have to attend graduate school for that!

Yeah I probably should. Don’t have the money tho.

I guess PKMS are only really good if you have an actual concrete goal in mind. “Learning advanced chemistry” is not a good goal. “Studying biophysical organic chemistry for medicinal applications” is a better goal, but still not specific enough to really be actionable and motivational enough for me. So what’s the point?

I probably need to reframe what PKMS are to me. I want to use it as a tool, but then how?

Let’s initially define PKMS to get this out of the way. A personal knowledge management system is a tool that helps clarify raw data (most often, qualitative/conceptual data) and crystallize it into digestible information. Said information then can be used as a reference or a tool to help me commit things to memory.

What do I mean by clarifying raw data? I mean that I want to gain new insights from said raw data by forming new connections that may not be obvious at first glance. I don’t know what would that even looks like as a process. Maybe there’s a way to tag raw data semantically, and related tags are shown in a better format so I can mentally connect things better. I don’t know.

Clarification of raw data into atomic notes is nice, but how do we elucidate said notes into a proper output, such as a blog post or a research paper? I would like a way to cite these atomic notes, but not in the same way I cite actual sources. Ideally these should be at least syntactically distinct. Such outputs preferably should bridge the gap between my sources of raw data, and the consolidated notes that I made in a way that distinguishes between what I collected and what I have synthesized. This bridge shall serve as the foundation of my intellectual outputs, marrying the raw data and the atomic notes.

Now that I’m writing this, it doesn’t seem to make sense to “bridge” these two together. Atomic notes are based around my raw data. It is a crystallized form of the collection of information from different sources. So saying the phrase “bridging the two” doesn’t make sense in hindsight. Maybe what I meant is that I want to separate the citation of my atomic notes (which should be hidden from the actual output) and my sources (which should be shown in the output as footnotes/bibliography).

flowchart LR
	rawdata[Raw Data] -->|visible| atomicnotes[Atomic Notes]
	atomicnotes -.->|hidden| output[Literary Output]
	rawdata --->|visible| output

Organization of digestible information in the form of atomic notes also seems vague to me. My initial thought is that it works like Wikipedia where each topic is its own note, and these notes are interconnected to other notes via hyperlinking. My problem is then the concept of being “digestible.” Wikipedia is not exactly digestible; it is comprehensive, however. Do I want that for my notes? I do want to refer back to it and actually re-understand it, so writing my notes in my own words would solve this problem. My other definition of “digestible” is more specific however. I want digestible information be easily memorize via techniques like spaced repetition. Better recall is one of my other goals. I have problems remembering things very well.

So uh. concretely. I want (1) atomic notes in the form of Wikipedia-style articles, (2) atomic notes in the form of Anki decks, and (3) only having one source of truth for both of them. Anki decks are of course for personal stuff (and maybe shared with close friends if wanted). The wiki articles are definitely going to be used for literary outputs/text projects/whatever you call them. I just wish I thought of a way to cite such articles in my projects in an internal way (for cross-referencing with my atomic notes) and an external way (via proper bibliography support).

Step 1 — Raw Data Extraction #

Organization of raw information is very hard, based on my experience in just starting out a new Obsidian vault. I’m currently using Zotero and the Obsidian web clippings extension, and even if it seems fine, I don’t like the added friction. Ideally:

1. I want to save my research articles and other stuff in Obsidian, including all of its metadata, main content, as text or PDFs if possible.
2. These should be easily cited, and citability should be as granular as per section of paragraphs instead of the whole note.
3. For PDFs, annotations should be easily cited as easily as block citing in Obsidian.

I have concerns regarding storing PDFs inside my vault. I would imagine that would bloat up the storage and be bad when I decide to store it on the cloud for syncing up stuff. That’s why I used Zotero as a separate storage system and annotation software for my papers. But then that adds friction to the process of extracting data.

The extensions for Zotero-Obsidian seems half-baked. I was expecting to use Zotero inside Obsidian, but instead they have importing annotations as its own function. I would like to have a hot-reloading system that auto-updates my annotation creation in Zotero to an Obsidian note.

Back then, I used an extension for annotation PDFs in university. I forgot which one. I’m not sure on how I would refer back to my annotations from my notes, because these annotations aren’t really linkable? I’ll have to revisit that.

If I have enough disposable income to pay for Obsidian Sync or any other synchronization services I probably wouldn’t have this problem.

Step 2 — Data Clarification #

Data clarification without a structured goal is hard.

Whenever I do a text project/literary output in university, writing things comes naturally to me. My process is as follows:

1. Write out the outline. It’s much, much easier to fill out the body text if you have a structured text project.
2. Search things up as I go along writing stuff in each section of the outline. Since I have a vague idea of what I want to write, searching is very easy to do.
3. Collect these sources, preferably in Zotero or in a Google Doc with a bunch of links. Then, copy-paste the relevant parts of these articles to paraphrase and synthesize in the actual text project.
4. After synthesizing text for each section, I then polish and organize the text.
5. Reread the whole project one or more times, polishing it a bit more for each pass.
6. Submit the text project to whoever needs it.

This is much harder when you don’t have a goal mind. What do you outline if you don’t have a goal?

In this case, the text project is now essentially a graph of notes that I can use as a wiki. So I guess the outline is now represented as a series of synthesized notes without any content. Which is what I’m doing right now. Now, I feel like that solution does not feel good for me because:

1. I’ll have to know what notes to scaffold. This is bad, since I don’t know what I don’t know. How should I even approach learning about new stuff in sciences? I am currently subscribed to RSS feeds of different journals, but these feel impenetrable to me. Should I subscribe to YouTube channels and newsletters aimed for non-scientists instead?
2. The volume of things I don’t know is very large. Every research article I read makes me add like fifteen new synthesis notes. Is that sustainable? I unfortunately did not build my Obsidian Vault from learning in university, and so I have not made notes about the things I’ve learned back then. I should go back with my books and create notes for them.
3. Continuing with the “very large” aspect of note taking, seeing empty literature notes is somewhat demotivating to me. I don’t know where to start describing the things I’ve learned in my own words. The context where I discovered these concepts are used in a higher level; they don’t really explain it to someone who is new to it.

Okay the previous bullet list should’ve been in the next section probably.

Have I actually explained what data clarification means to me?

For me, I would like to understand each source I get by rephrasing it in my own words. Cornell notes style of having a summary section. This also includes asking questions about the source material (or basically lectures in their case). So I guess one way to clarify data is to ask questions about it first, and then answer them. That’s one way of paraphrasing.

But then how do you know if your questions are good?

This is not gonna be measurable. Like, I’ll have to just keep going and do a retrospective to judge whether or not my questions are good. I guess that’s fine. My wiki is just gonna be burdened by the past, especially since I don’t have a git repo set up.

Anyways. Data clarification. How do I make things clearer to me? How do I elucidate properly? I don’t have the answers yet.

No, I’m not gonna use AI for extracting and clarification of data. However, the thought of setting up my own local LLM model that runs on my literature notes seems interesting.

Step 3 — Literary Output #

Time for the actual writing!

Uh.

For basically a monthly newsletter you can’t really outline, do you?

Outlining is very easy if you have a planned text project, like blog posts and theses and review articles and whatnot. Monthly articles less so. I barely have any experience in newswriting, so this is gonna be hard for me.

But honestly, the previous few steps are much harder than this one. It’s easy to get into a flow state of writing as long as you have an outline and a vague semblance of an idea you want your text project to be about. If you also have a criteria to follow, that’s fun! That’s much easier.

There’s nothing much I can say about Obsidian here. It’s a markdown editing file. I wish I could use my MDX components, or do some theming shenanigans, but that’s honestly just overkill. I can understand the concern of having more plugins means that it’s harder to migrate or maintain your vault. I just wish Obsidian got more inbuilt features.

Speaking of, the way I type my blog posts nowadays is so weird. I write it in Obsidian, with each outline point as a separate note. I guess it’s fine. I do not like to scroll down too much when browsing my drafts.

Putting my content into the blog site is just very manual. A simple copy and paste is just done, plus some fiddling with the metadata stuff and uh. git push. Yup.

I don’t really know any way to do it better. Also, I’d also like to use Typst for publishing, but I have not seen plugins related to that.

So I guess my wishlist for Christmas is a way for Obsidian to export to

1. Astro via MDX files (I’m not setting up a CMS. I have custom made components)
2. Typst
3. Okay I change my mind. Is there a CMS that allows custom MDX components/tags?

I guess I’ll have to make an update post when I get to grow my Obsidian Vault.

Conclusion #

Uh oh we got into the concluding part of this mess of a blog post that was in draft hell for six months!

Anyways. I hope this blog would reach people who are more familiar with Obsidian, and people who are familiar with the Zettelkasten system. I want to learn more about these things, and I would love to make these tools fit better with how I work.

Obsidian is a nice software, but I’ll probably want to have something more specialized than this. I realize that I like wiki-style software like DokuWiki and MediaWiki, but I don’t like that it has to be done as a server or be interfaced via a browser.

I don’t like how solutions to turn Obsidian into a wiki requires fiddling with CSS, instead of I don’t know, plugins that add more syntax to markdown? I’d rather have that, to be quite honest.

This article is admittedly a mess but I don’t mind. I guess this is just a six month decompression blog post in the making because I’m kinda overworked and underpaid and I want to have a better environment for myself to write something out.

Addendum 1 — Pitfalls of Obsession with Prescribed Systems #

One thing that’s putting me off about Obsidian is when I read this blog post.

Please read the article, especially if you are into note-taking and productivity software.

It made me rethink about the amount of videos I’ve watched and articles I’ve read about the different systems we have. And I’ve realized:

I don’t really need to follow these stuff to the T.

Like sure. I think Zettelkasten is useful, and I’m currently trying out a modified way of doing that. But I don’t think everything can fit neatly to atomic notes, and not all projects lend well to that type of knowledge management systems. I’m recognizing that a wiki-style format is preferable to my tastes, but that kind of system doesn’t lend to step 1 and 2 of my idealized process. The synthesis notes would then serve as basically the wiki pages and consequently a text project upon itself. Do I want to take the effort to do that? How exactly should I format this synthesis page anyway?

Also, this article is a reminder that these methods and systems are just tools. I do not have to adhere to them 100%, nor do I have to be shackled. Well, not like that. I guess I just don’t want to be too into the tools I have discovered, but then having no system is just not helping me at all.

Addendum 2 — Digital Gardens #

The idea of digital gardens is very appealing to me. Yeah sure, not being confined to time-based submission of text projects is very tempting. Not having the expectation of having a polished web project, but rather a public display of a set collection of unfinished notes in my Obsidian vault. Problems with this:

1. I use Astro for my site, and I’m not sure of a good way of integrating it with Obsidian.
2. I copy paste from sources to my notes, so publishing those would probably lead me to trouble with plagiarism stuff.
3. I like to do blog posts!

Some posts I’ve read about digital gardens are telling me that it is a replacement to the chronological aspect of the internet. Sure, I get the premise of the argument. Back then, web pages were curated and chosen carefully for the people to see. There’s a topic for cooking, some for personal stuff, some domain specific knowledge, et cetera, et cetera. Well. this blog is titled “ramblings for @harudagondi”! I’m not really organizing much here! lol

I’d rather have a tag based system with some better hyperlinking, not a section-based website like most of the digital gardens I’ve seen. Maybe if I was a niche website, like one of my favorites: Mountain of Ink. But I’m not that. I’m messier. Digital gardens imply a sense of thought in my website organization. Well. I’m tagging this post and leaving it at that. Sorry.

I don’t think blogging broke the internet. Rather, I think that the ease of making a blog diluted the specialty of the internet, and monetization worsened it. Drats. Capitalism is at it again.

Maybe I agree that chronology made the net a bit bad. But I think it could be easily resolved with a better tagging system. I don’t want to “cultivate” a digital garden. That takes effort. I’m not looking to earn twenty billion dollars from this, nor perform a sense of “thoughtfulness” and “meditation” about the life I live in the digital space. No thank you. I just love to write, and I want some people to read it. Go figure.

I think blogging is just the easiest of doing that without overthinking what my digital garden should look like. Despite people saying “digital garden” is a less performative way of posting, I just probably have it mentally categorized as “how to perform being messy.” Maybe I should talk to a therapist about that.