In reply to On taking notes and syndicating them by zblesk

Thank you for writing that! I too find it highly interesting to see how other people arrange their workflows, choose their tools and what they do with them. Often there are things that spark an idea or suggest a useful tweak to my own workflows. So thank you for making a comparison between how you work and what I wrote about how I work.

A few reactions to some of the things you mention.

My perspective on (personal) knowledge management is centered around the notion that I should have everything under my fingertips, and should be able to fully determine my own choice of tools. Tools one can preferably tweak, reshape or replace easily. I started taking notes in the early 1990’s and local text files were the most basic choice I made (and one of the few I then could make). Later convenience lured me into other things like Evernote, and Things for tasks, but I’ve returned to that basic starting point of using text files more recently.

At the core are these notions I hold:

  • Local first. I’m from an era that connectivity can’t be taken for granted, and regularly work in settings where that is still true. It is also a dependency that even when it is usually reliable, probably carries a high cost if it does fail, as that most likely is in key moments (basically a version of the demo effect).
  • Agency over tools. Tools must provide actual agency. A key part of it is being able to fully control it’s deployment and use, being able to tweak it etc. Tools must be smaller than us in that sense (not in a literal sense). Convenience may make me ignore this factor up to a certain point, but in the end having control over my tools always comes back up as an issue. Not having such control ultimately always turns a tool into a single point of failure. (Gmail and Evernote are prime examples to me) That drives me to simpler tools within my own scope of control and power to manipulate, and only allowing more complexity if it increases my personal agency significantly. It also means to me that tools need to be useful on their own, and more useful when networked.
  • Personal tools. Tools need to be adaptable to the person using it. That makes it easier to make those tools smarter. As personal preferences can be assumed as the defaults, and personal routines are predictable to the person itself. Predictable routines plus preferences equal functions and parameters, i.e. code.
  • Personal agency is always in the context of networked agency. In most settings the unit of agency is not the individual but a small group of connected people trying to solve something that is important to the group itself. Whatever tools the group uses should be within the scope of control of that same group. As a group’s notion of local is usually a networked notion, my local stuff needs to be able to connect (yet not depend on it). Distribution is important here. Centralisation is mostly to be avoided as it carries a cost in overhead, control and resilience.

Put that all together, and indeed POSSE basically becomes the prime directive for everything.

On PHP: I’ve been using PHP for about 20 years. When I create something as a personal tool, it makes sense to just grab the building blocks I’m already familiar with. I’ve always run a local webserver on my pc/laptop, and writing up a few lines of PHP and dropping it in a folder my local webserver uses, makes it fast and easy. Easier at least than getting to know new frameworks, or whatnot. Javascript never appealed to me even if it is from the same 1990’s era, nor succeeded in making much sense to me. Apart from doing browser side things with it in an HTML page.

On WordPress: I used to handcode my sites, until I started using Movable Type shortly after I initiated my blog (hosted on a webserver at home). That was written in Perl which I was comfortable with having written my then employer’s first intranet in it. A decade later I switched to WordPress when my Movable Type install suddenly stopped working completely. I see you use Ghost, which ran a kickstarter I supported shortly after I switched to WordPress (self hosted on an external hosting package). By the time Ghost saw its first release I didn’t act on my earlier idea of running that on a home server. I’m not particularly attached to WP (also used Drupal heavily for other sites), and use it pretty bare bones, but it has served me well for the last 10 years. The switch to Gutenberg and blocks though has me thinking I might maybe go for something simpler.

On Obsidian / Joplin: I also use Joplin, but haven’t tweaked it like you have, I use it out of the box. It’s where my Evernote exports live, which from there I export to md files as needed. I treat Obsidian as a viewer, and Joplin too. Because of that I dislike that Joplin stores stuff locally in an sqlite database, obscuring the contents from my filesystem that way. From a viewer it then becomes an obscurer. Currently Obsidian has my sympathy, that may change, no tool is forever. So in my choices of e.g. plugins for Obsidian I avoid things that provide functionality that comes with a type of lock-in, where if you stop using a plugin part of your information disappears or is hard to get at because it was in a database not in the notes. I dislike YAML frontmatter too. For the Dataview plugin I use inline datafields (key:: value) which makes them a regular part of the note itself. Only when for some automation I need to know where to easily find a data field, I will put it at the top (but still not as YAML frontmatter).

On public RSS subscriptions: yes, I post a list of all the feeds I subscribe to. I treat them as individual’s voices (so no feeds from news outlets etc), and group them by my perceived social distance. I treat blogging and interacting with feeds as distributed conversations.

I always like reading about how other people process information and handle their notes/knowledge bases. It’s a topic I think about often.

Ton Zijlstra’s ideas are especially interesting to me because it seems we are trying to achieve similar goals, but go about it in opposite ways.


(also posted to Indienews)

In het past weeks I struggled to get to action. I didn’t have the sense that I was in the pilot seat. Too many little things budding in, not being able to get started on bigger things, and no sense of overview. Or rather, a too overwhelming overview, and no easy way for myself to bring the scope of that view down to something manageable for the day.

I have about 35 areas of activity, this includes projects, general tasks, both business, volunteer work and personal. For each of those 35 or so activities I keep a running list of things to do. Some lists have a few items, others have a few dozen. If on average they hold 10 tasks each, it means a tasklist of 300 to 400 items to choose from. That makes for an overwhelming overview. It gets better if I dive into the scope of a specific project or activity area, but then I don’t see the small things I can do to keep the other activities rolling. When I was still using Things I had the same effect, so this isn’t something particular to my current use of Obsidian for tasks.

The result has been that, because the overall list is too overwhelming, I haven’t been using it much. Which means I have even less sense of overview or being on top of my stuff.

In an attempt to regain that sense, I’m now trying to each morning go through the entire list and pick a handful of things for the day. That small curated list has a more manageable scope, without being limited to a single activity.

I don’t want to copy tasks from one place to another. I want them to stay on their respective project or activity list, but marked and summarised on my daily list. I’m aware there are various task oriented plugins for Obsidian, but they will prescribe me a certain mode of working, and it isn’t certain that in their absence the same information will be as usable / findable (a type of functional lock-in or dependency I always want to avoid)

What I came up with is I mark every task I choose for the day with ‘t::’, in whichever line of whichever file I want. This can be an existing tasklist, but I can do the same while making meeting notes, to quickly mark something as a task resulting from the meeting. The Dataview plugin I already use sees ‘t::’ as an inline datafield and is able to extract them into a list using the following brief piece of code:

TABLE t as Vandaag
SORT File asc

I display that at the top of my daily note. It allows me to quickly jump into a task list or other note to delete it when done, and to copy it over into my daily note in the ‘done’ list.

In the coming days I will test if this improves my days and activities.

A brief list of selected tasks from other files. Also note that at the top t:: is mentioned inline twice, and both show up as task items in the list.

Bookmarked Research Rabbit

Research Rabbit is a tool that, when provided with some academic paper you already are familiar with, can suggest other related material as well as provide that material. By looking for material from the same authors, by following the references, and by looking at the topics. This can speed up the discovery phase quite a lot I think. (And potentially also further increases the amount of stuff you haven’t looked at but which sounds relevant, thus feeding the collector’s fallacy.).

I’ve created an account. It can connect to Zotero where you already have your library of papers you are interested in (if you use Zotero with an account. I use Zotero standalone at the moment I added a Zotero account and storage subscription to sync with Research Rabbit).

Looks very useful. HT to Chris Aldrich for in Hypothesis pointing to a blogpost by Dan Alloso which mentioned Research Rabbit.

Op 20 november a.s. vanaf 20:00 vindt de derde Nederlandstalige Obsidian meet-up plaats! Dit keer geïnitieerd door @CABenstein in het Nederlandstalige kanaal op de Obsidian Discord. Je kunt je aanmelden op Eventbrite, of anders laat even van je horen op Discord.

Tijdens de sessie is er alle tijd om tips en tricks uit te wisselen over het werken met Obsidian. Zelf ben ik altijd erg geïnteresseerd hoe het persoonlijk kennismanagementproces (pkm posts op dit blog) van mensen is georganiseerd, en hoe ze dat in hun tools vormgeven. Over de eerste meet-up schreef ik een impressie, en dat deden Wouter en Frank ook.

Schuif aan!

Amazingly useful plugins keep getting made for Obsidian. Plugins that help reduce friction to getting my material from other sources into flat markdown files that I then can edit, rework and do with as I see fit. Earlier I mentioned the Zotero plugins to extract PDF highlights and Zotero links into Obsidian. Today I started using the Kindle highlights plugin. It connects either to your Amazon account or you feed it your myclippings.txt file from your Kindle device.

It then pulls in your highlights into one note file for each book. You can edit the template for that note, to make sure it includes the information and metadata you want. Each highlight has a link directly to the highlight in your book, and when I click it opens my Kindle app on my Mac and jumps right to it.

Previously I would either download a CSV through the export notes function on a Kindle device which mails you your notes, or copy the myclippings.txt.

Using myclippings.txt is still the only way to get highlights from a book you did not add to a Kindle through Amazon (e.g. a direct upload from my Calibre library).

With one click, upon my first sync I now have about 100 notes with highlights from books I read. I never was a heavy highlighter, because of the friction of getting those highlights to a place where I could use them. That may now change.

Image: a screenshot of how highlights from one of my books now show up as notes in a markdown file, using the default template.

De allereerste Nederlandstalige meet-up van gebruikers was interessant en leuk! We waren met z’n vieren, Sebastiaan, Wouter, Frank en ik, en spraken bijna 2 uur met elkaar. Leuk om te vergelijken waarom en hoe we notities maken in Obsidian, te horen over de (vaak lange) historie die iedereen heeft met notities schrijven, met de hand of in een reeks tools. Het gesprek ging vooral over onze werkwijzen bij het toevoegen van notities, en in mindere mate over het gebruik van die notities.

Dat gebruik is zeker iets om over door te praten een volgende keer, en voor mezelf om bewuster bij stil te staan. Ik merk dat ik nog vooral in de modus zit om mijn collectie voldoende te laten groeien, en minder in het bewust raadplegen van mijn materiaal als ik met een onderwerp aan de slag ga, of iets wil maken. Tegelijkertijd merk ik wel dat ik nieuwe verbindingen leg tussen notities, en (omdat ik mijn oude weblog postings tot notities verwerk) tussen oudere en nieuwere gedachten waarover ik heb geblogd. En dat ik in het afgelopen half jaar een paar blogpostings en twee presentaties heb gemaakt die direct voortkomen uit het samenbrengen van bestaande conceptuele notities. Dat soort productieve uitkomsten is al heel nuttig, maar ik ben nog op zoek naar het een grotere rol geven van mijn notities in het reflecteren op een onderwerp, in het nadenken over vragen etc. Dit allemaal ten aanzien van conceptuele dingen dan, want m.b.t. mijn uitvoerend werk zijn mijn notities het werk als het ware.

Daarvoor is in ieder geval regelmatige blootstelling aan mijn notities nodig denk ik. Enerzijds door beter te weten wat ik er in heb zitten, en anderzijds door me in mijn workflow vaker bezig te houden met wat ik over een thema al heb bedacht, verzameld en geschreven. In de afgelopen weken ben ik daarom meer index-achtige notities gaan maken, emergente outlines. Door anderen wel maps of content genoemd, ik noem ze olifantenpaadjes. Zo’n notitie verwijst vooral naar bestaande notities, met wat zinnen er bij waarom ik die links bij elkaar plaats. Zo’n verwijs-notitie helpt me makkelijker de weg te vinden in mijn digitale tuin. Ik maakte eerder al emergente outlines maar die waren niet op navigatie gericht maar op het samenbrengen van notities in een ruwe verhaallijn, dus meer op schrijven gericht.
Een andere manier van blootstelling is via spaced repetition de inhoud van mijn notities voorbij laten komen, in de vorm van Anki vragen bijvoorbeeld. Sebastiaan liet gisteren zien dat hij een plugin aan het maken is die dat binnen Obsidian zelf doet, niet gericht op vragen of onthouden, maar gericht op herlezen.

De belangrijkste ingang voor hergebruik van mijn eigen materiaal is vaker mijn notities raadplegen als ik met een vraag aan de slag ga. Te veel leun ik op mijn eigen actieve herinnering, en dat leidt vaak tot reconstructie van dingen terwijl ik die al keurig heb uitgewerkt in mijn notities.

Wouter schrijft veel met de hand, en liet zien hoe hij al die handgeschreven pagina’s scant en in zijn Obsidian vault opneemt. Op die manier kan hij ze raadplegen en er naar verwijzen als hij er dingen uit tilt en in een notitie opneemt. Dit is een interessante manier. Al schrijf ik veel direct digitaal, ik maak ook veel handmatige aantekeningen, en er ligt hier een stapel A5 notitieboekjes. Nu heb ik ook een staande scanner met voetpedaal van CZUR, en die leent zich goed om snel notitieboekjes te digitaliseren realiseerde ik me na de meet-up. Daar maar eens mee experimenteren.

Tot slot, was ik blij met hoe prettig mijn Jitsi server werkte zo met z’n vieren. Benieuwd hoe dat gaat als je met een grotere groep bent. Alleen het scherm delen had een rare bug waarbij de bovenste strook van een scherm bleef hangen op het eerste beeld.

Wil je ook met andere Obsidian gebruikers in het Nederlandse praten over je ervaringen? Kom dan naar het #nederlands kanaal op de Obsidian Discord server.