« Simulated Annealing

Deconstructing digital gardens

How we can build better knowledge systems

“Digital gardens” are a new take on presenting information on blogs. Rather than a linear series of blog posts, these digital gardens are instead a networked collection of ever-evolving notes. Built into the concept is the idea of “growth” – there is an upfront expectation that the notes and the graph will change over time. The content is an ongoing work in progress, with the information dynamically being recombined, edited, or removed at will.

People who have adopted them seem to find a lot of value from this new, emerging interaction paradigm. The case for digital gardens can be boiled down to a few principles:

  • Rejection of perfectionism and the oppressiveness of the “essay”
  • Creating new digital interfaces
  • Disdain of “performative” writing
  • Exploring ideas publically (“thinking out loud”)


garden_example Mental Nodes is a great example of a digital garden.

I have considered adding a digital garden to my own website and decided against it for several reasons. I wanted to break down the various aspects to weigh the positives against the negatives, as I see them.

Digital gardens have the major benefit of lowering the friction to write and share ideas. Because there is not an expectation of a polished essay, writers can quickly capture good ideas and connect them to other concepts. Each note will usually have multiple hyperlinks to other notes, and off-the-shelf libraries like TiddlyWiki can make it easy to add new links and connect your notes together.

From a reader’s perspective, digital gardens are much better than a blob of unindexed notes. The lightweight, loose structure of linked notes make it easy to traverse the content graph, exploring related ideas.

I do question whether digital gardens are inherently superior to essays. I can understand the dislike of formalism – there is a lot of pressure and work that goes into producing an essay. Removing the expectation of perfection makes it easier to get started and just write something. The essay format also forces writers to come up with a linear structure that may not be well-suited to an exploration of ideas. And once it is published, written essays are typically “done” and never changed. This can make it hard to write on topics where your position is not fixed or might frequently change.

For me, the benefits of the essay outweigh the implicit expectations or pressures that come with the format. Writing an essay forces me to organize my thoughts into a coherent structure and clarify my arguments. Not only does creating an essay improve my writing, it makes me a better thinker.

If overcoming the mental burden of writing formalized “post” is the main issue, I prefer the approach of Justin Duke. He compares his personal site to clay and references other similar writers who are rebelling against the staticity and doneness for conventional written pieces. Their sites have published posts with statuses like “in progress” or with empty, stubbed out sections. Brian Lovin iterates towards an “incrementally correct” personal website by organizing his work by updated date rather that a single publication date. All these techniques help lower the personal pressure for perfectionism, while still using the traditional essay form to structure their thinking.


meta_tags Gwern uses multiple meta tags to denote the progress, certainty, and importance of specific written pieces.

I believe there is a ton of value in networked notes generally. I am not surprised that the increase in digital gardens in the blogosphere is closely mirrored (or maybe driven) by the rise of new “networked thought” software like Roam Research, Obsidian, and Notion. I use Obsidian myself for my note-taking system and it’s been incredibly valuable for capturing and exploring ideas. The seeds for this essay germinated in the fertile soil of my Obsidian notes.

But in my opinion, where digital gardens (and networked thought systems, more generally) fall down is in the reader experience. An essay can involve extremely complex concepts but necessarily must be understandable to be successful. The format itself forces the writer to tell one complete argument or story in a linear fashion.

In contrast, it’s hard to know how much of the content graph to explore for a digital garden. Since every note can have multiple branch points to other notes, readers need to repeatedly backtrack or open up multiple tabs to get a complete picture of the related ideas. There’s also the problem of entry and exit points. Because there is no requirement for a start to a networked graph, it’s hard to know where to start reading. And once you start, it’s unclear where to stop.

My general experience with digital gardens has gone something like this: I encounter a particular note in a digital garden from a writer. The concept is interesting and I want to read more. I then spend the next five or so minutes randomly clicking around the networked notes, until I am bored or distracted by other things. The time spent in the digital garden also decreases dramatically if the first few links I click on are deadends, empty, or not very interesting compared to the original note. I usually leave the website feeling unsatisfied and worried I might have missed something important.


wiki Even Tom Critchlow’s digital garden, which has a lot of useful entry points and meta structure, is hard for me to read thoroughly.

That’s why I think these networked thought tools really shine on a personal level, for the writer themselves. When you are already familiar with the shape of the content graph, you can quickly explore the concepts and follow interesting trains of thought. It becomes possible to evolve and shape your thinking, especially by connecting disparate ideas that were not originally connected. For an outsider who doesn’t have the super-structure already imported into their brains, a digital garden can feel like a chaotic mess.


graph A snapshot of my “second brain”, my densely-networked graph of notes. Hard to know where to start, especially if you’re not me.

It’s especially frustrating for certain types of readers. I personally read very quickly and am constantly overloaded with information. My reading backlog is full of hours and hours of articles, papers, and essays. I like to heavily skim any new piece of content to see if it’s worth a deeper read. I can get a quick overview of the important points and only then add it to my ever-increasing queue. It’s essentially impossible to quickly skim a digital garden because the interface of the web gets in the way.

In it’s current iteration, digital gardens take active effort on the part of the reader to wade through links to related pieces of content. This is very unlike a real garden: you don’t have to be an expert at horticulture or garden design to appreciate the overall landscape. You can easily engage with many different levels of a real garden at the same time, telescoping in and out at will to first examine the minutae of informational placards or specific plants and then returning to a broad perspective of the scenery as a whole.

That’s not to say digital gardens will never overcome the awkwardness of the reading or “viewing” experience. I actually think there’s immense potential in the direction of digital gardens. From the beginning, digital gardens emerged as explorations of new ways to interact on the web. It should be possible to create elegant, automated solutions to make it effortless and enjoyable to consume content within digital gardens.

To perhaps overuse the garden metaphor, just as the best real gardens have consciously placed walkways that lead you to the most beautiful vantage points, so too should digital gardens have guided paths through the content.

A great example of a guided experience can be found in the Google Arts & Culture interactive experiment Ocean of Books, an abstract map visualization of the entire space of literature.


ocean_books Ocean of Books places authors on the map by their proximity on the web, based on how frequently they are mentioned together.

You are able to explore the map in a free-form fashion, zooming into particular genres and authors at will. Just like with digital gardens, I quickly found myself running out of steam once I engaged with a few authors. The vastness of the space to explore felt overwhelming. But rather than leaving people to navigate the space entirely on their own, Ocean of Books added optional hot spots which, when clicked, lead to different points on the map. The guided points both offload the decision-making of where to explore next and also took me to sections of the map I was unlikely to explore on my own.


hot_spots An example of a hotspot on the map. Tapping on “Next” will take you to another author, sometimes in a completely different area of the map.

While the Ocean of Books guided path is curated and always surfaces the same hot spots each time, I can imagine a dynamic system for generating guides for readers to explore your digital gardens. These guides could factor in various inputs, such as last updated time, the size of the note (“more finished”), time spent reading, or how close or far apart two points are in the graph.

I’m excited for more knowledge tools that can easily enhance the reading and content discovery experience for website readers. Toby Shorin is one of the few people building tools for a better textual screen experience. I especially like Open Transclude, a simple UX plugin to add rich networked quotes within blog posts. The goal of Open Transclude, and many of Shorin’s other UI experiments, is to provide meaningful and rich context to his other writing without letting the reader get lost in the morass of the entire body of his work.

Shorin articulates the same tension between structured essays and dynamic, interlinked knowledge graphs:

Knowledge is not an accumulation of facts, nor is it even a set of facts and their relations. Facts are only rendered meaningful within narratives, and the single-page document is a format very conducive to narrative structure. The hypertext books that have gained popularity (I’m thinking here of Meaningness.com) have largely conformed to this in two ways: 1) there is an intended reading order, and 2) the longer essays within the project do most of the heavy lifting in terms of imparting the author’s perspective to readers.

On the other hand, the notion of the “document” that is intrinsic to web development today is overdetermined by the legacy of print media. The web document is a static, finished artifact that does not bring in dynamic data. This is strange because it lives on a medium that is alive, networked, and dynamic, a medium which we increasingly understand more as a space than a thing.

Another area that needs to be solved before digital gardens can be a truly rich ecosystem for readers is content distribution. Right now, email newsletters and RSS feeds are enjoying a new revitalization as digital content consumers move away from the endless streams of social media. However newsletters and RSS are built for the fundamentally atomic, chronologically increasing format of the traditional blog post. Relying on people to remember and constantly check for newly updated notes in a digital garden is not sustainable. I imagine there is a whole class of tools that can be designed to automatically capture and format coherent segments of the knowledge graph to push to readers.

Again, there are wonderful people on the internet re-designing mainstream reading experiences. Many of the approaches don’t even necessitate building new software or tooling, and are simply upending established convention. I love Julia Evans’ automated weekly mailing list which sends you a fixed list of emails on a weekly basis, rather than the “latest” post or update. This could be one possible approach to a digital garden newsletter, where the writer crafts a fixed series of emails based on different core concepts in the knowledge graph.

I think we’re on the frontier of a whole new way of exploring knowledge and deeply reading on the web. It’s no surprise to me that many writers who are also software engineers have co-opted Github to write essays, personal statements, or even structure their entire digital gardens. By using software built for programming, these writers get change detection and revision history on their own work for free. Any reader can also explore how the writing has evolved over time. How will reading be different when all these capabilities are offered as a first-class integrations in publishing technology?


There are only two arguments for digital gardens that I find unconvincing. Those are (1) the disdain for “performative” writing, and (2) the strict positives of “thinking in public.” I think the disagreement stems from a difference in perspective. These points are typically being made by writers with very large followings and I, with my very nonexistent following, have different pressures and incentives.

I find it strange that blog posts are now categorized as “performative” writing. When I write, I’m doing so to exercise my own reasoning and practice the craft of writing, and maybe save some readers the trouble of doing the research to get to the same conclusion. I consider my blog to be in the grand tradition of the bedazzled Geocities websites of the early 2000s: unrefined, mostly a labor of love, and hopefully entertaining to the few unlucky individuals who happen to stumble across it.

I can see how blogging could transform into “performance” as writers reach huge scale with their audience. They may feel more pressure to write on topics that generate a lot of traffic and engagement, rather than those that are truly interesting to them. That, plus the added formalism of extensively editing an essay to perfection, can feel very artificial and constrained.

I’m not sure digital garden is necessarily “less performative.” If anything, the publishing of every little unfinished thought seems more performative to me than constructing a well-reasoned narrative. I’m personally very skeptical my random notes are interesting or useful, so it feels excessively presumptious to assume that others want to read them. As Jia Tolentino writes in her excellent essay collection, Trick Mirror:

To try and write online, more specifically, is to operate on a set of assumptions that are already dubious when limited to writers and even more questionable when turned into a categorical imperative for everyone on the internet: the assumption that speech has an impact, that it’s something like action; the assumption that it’s fine or helpful or even ideal to be constantly writing down what you think.

But again, it’s a difference of scale and perspective. Writers with a large audiences may truly be providing a lot of value by sharing their notes.

As for the second point, I can see the argument for sharing unfinished ideas early. By “thinking out loud”, others may find interesting concepts and start a productive dialogue with you. I don’t think these occurrences are necessarily limited to writers with thousands of followers, since the beauty of the internet is that anyone’s writing can be discoverable.

But what does limit the possibility of discussion and brainstorming via digital gardens is, again, the suboptimal reader experience. Because I have no way to easily include digital garden content in my newsletters or help random passerbys traverse through my knowledge graph, the only ones who are motivated to dig through a digital garden are people who already know they like the content. These fans are probably the ones starting discussion threads on Twitter, creating that enriching experience for both sides. Small-time and hobbyist writers will likely not see much interaction because of the friction involved to exploring digital gardens.

Public digital gardens in their current form are only really beneficial for established writers with large followings. But there is so much potential to transform how we read and engage with text on the web. I can see a future where readers can encounter and truly traverse digital gardens, in all different flavors and sizes.


garden


Appendix

If you enjoyed reading about my thoughts on digital gardens and want to explore some yourself, check these out:

On working in public:


This article was last updated on 09/18/2022. v1 is 2,654 words and took 7.9 hours to write and edit.