Digital Gardens

Some of you probably know that I generally oppose the idea of public/collaborative “notes” intertwined with personal ones. I believe in a difference between language (voice?) used in personal notes (as a tool for thinking), vs public notes (as a way of communicating).

At the same time I love the idea of working with the garage door open, concept mentioned here a couple of times.

Let’s collect public digital gardens, some that I know of:


A few months ago I was experimenting with creating a digital garden on my own site though I definitely need to update it with new and updated notes.

Something I was experimenting with was what sort of information I could convey about a page before someone clicks on it. Here I am generating Chlandi frequency patterns based off the word count (length) of each page.

I would also suggest to take a look at some of the wikis on

1 Like

I’ve been talking to a few people about this for a while, but is it worth talking about a product that would let people easily build their own digital gardens? Some of these ideas are covered in my Information Forest post as a holistic note-taking browser, but I think this can also be a web app. Everyone has their own personal touches which is awesome, but the barrier to entry is quite high to make one of these.

Off the top of my head, the best features of all of them are:

  • Ability to make curated lists / collections
  • Tagging
  • Backlinks
  • References
  • Rich previews

And where these can be improved:

  • Forking / remixing / embedding
    – This is more complicated and deserves more design time, but being able to reference a portion of a garden in your own (like a quote tweet), or fork and asynchronously pull in changes at will
  • Version control / history with branching
  • Auto-tagging, summarization, keyword extraction, named entity recognition (all NLP stuff)
  • Better navigation (a la losssless spatial navigation)
  • Full text search

Some crazier thoughts:

  • What does “merging” two gardens look like? Wouldn’t it be so interesting to take two people’s digital gardens and see overlaps? Primary differences? Where they fill in the blanks? Imagine seeing that if I’ve saved two links, another person saved a link that bridges those two! What about the “collective garden” of 1000 people all interested in the same thing? Different things? Teams? Organizations?
  • What about an API that lets you query a digital garden from any other client?

Some other thoughts:

  • How can we incorporate end user programming into the platform? Maybe make it API first?
  • How important is the customized visual design of each garden? Should visual customization be a top priority? (seems like it might be. at least on the level of basic color palette + font choice).

So, some of these problems are primarily NLP/ML problems, and some of them are primarily complex design problems. But I think they’re all ultimately extensions of link saving/sharing with

What do you all think?


Also, here is a collection of digital gardens I’ve found so far:

1 Like

I’ve been talking to a few people about this for a while, but is it worth talking about a product that would let people easily build their own digital gardens? Some of these ideas are covered in my Information Forest post as a holistic note-taking browser, but I think this can also be a web app.

For me, web apps don’t seem like a good solution. I don’t think that web apps will ever by as reliable and fast as local files, they also tend to disappear. Thinking-notes seem to make the most sense if I can hold on to them throughout lifetime.
Viewing and editing plain text files seems probable even in 5/10/50 years, not so much for accessing online database of some service (even self hosted).

I agree though that the barrier of entry seems high right now, but I’d prefer to look at it through the lens of simple composable tools.

I currently explore small steps in this direction with my muninn stack: + though it seems far from general user accessibility.

I think that these tools should operate (idempotently) on plain files/folders, providing custom materialised views. This way users don’t have to worry about loosing/overwriting or editing their data.

Then we wouldn’t need to worry about end-user programming for this system really, since it would remain editable by any script that can read/write plain text files.

And then, to create a hosted digital-garden from those notes, one could run (one of multiple) static site generators, maybe with some custom addons - what do you think?


interesting. this is definitely one approach to the problem-- and I think it would work for a lot of developers and is the current pattern many are following with their own digital gardens. in fact this is probably where I’ll start (local tools + custom scripts).

I really like the idea of each component of this system being like its own app that can exist (in your words) idempotently. loosely coupled and compatible with any system that deals with plain text.

but I don’t think it’s incompatible with webapps. maybe the key is making any data you store on the webapp completely backwards compatible with your local file system-- and the webapp is simply used as a social layer on top of your garden. for example, all the web app would do is sync to your local file system, create a custom view layer on top, and allow you to access other peoples’ gardens if they’re published. so you could see this as a local-first, low-code garden builder, composed of several utilities (like your various plugins) that can easily be hosted by a central service. this way you’d be able to easily self host as well. but unless this was P2P (which adds another layer of complexity), I really want to be able to follow other people, subscribe to gardens, etc.

the dream im imagining is that, given a standard protocol for a garden data format, you could use any tools you want locally and it would automatically be consumable by some garden-renderer. EDN, which is what Roam uses I think, seems like a good candidate for a hybrid data-store and data format for a garden.

so, given the constraints that we want:


  • based primarily around plain text (or easily serializable, human readable data)
  • easily composable
  • local first and scriptable

bonus points:

  • can be shared easily via URL / hosted / deployed
  • collaborative (ability to interact with other users if they publish), subscriptions, etc

what do you think are some good approaches? It seems like social features should be just like another plugin that gives you access to publish/subscribe/comment, but everything else can remain the same.

1 Like

I did this during my PhD thesis, although they were very structured around literature notes.

I have been blogging for 14 years, and now I am experimenting with sharing some of my notes from Roam (I have a workflow going from a Roam JSON export to a Gatsby static site). It’s very helpful for being able to share specific things with people… It would be better to spend a lot of time writing it up for an audience, but in many cases the decision is between sharing something unfinished or not sharing at all.

One interesting usage of private sites is the course I took in 2007 with David Wiley, where everyone were blogging on their own sites (I realize blogs are somewhat different from digital gardens, but the principle is similar). Because everyone were writing each week on their own spaces, I still have my posts that I wrote then, 14 years later - and the course served to kind of bootstrap my personal learning network - whereas traditional courses take place in a specific platform and disappear once the platform closes. (Even this discussion - if the website went down, all of our thoughts would be gone… What if we could have written this on our own digital spaces, and easily merged them into a thread)?

So very interested in the process of going from individual to shared notes, but also on how we can enable more collaboration, interlinking, notifications, being able to see others who have written notes on the same article, etc.

1 Like

I believe in that difference too. But I also see it as a gradient from thinking in your own words to gradually remove context so that it becomes easier for others to follow your thought process.

As such I find Digital Gardens a fascinating solution to fill the gap between personal note taking tools and internet publishing.

I’ve been looking for a way to gradually turn my notes into versions that others can understand, without going through a full-on writing process where I start with outlining a full narrative before I publish.

In my experiments with this, I came across “blog chains” as one fascinating idea that also fills that gap ( It does feel a little closer to the publishing side of the spectrum to me. Digital Gardens on the other hand seem to be closer to the note taking side.

Has anyone found a Digital Garden setup that’s easy to get started with and doesn’t require too much technical maintenance? When I glanced over what people use it’s either proprietary or very heavily customized — which is fine; it’s a freshly evolving space and that is how it starts.

I came to the conclusion that such an issue could be addressed firstly by a novel literary style.

Narratives, essays, etc. they are all very common styles in which we tend to format our thoughts. What these styles all have in common in that they are very scholar in nature.

Such styles may not seem like it, but they are very important tools. So, if a literary style is a tool, could there be some other literary style that is more liberating, accessible, and versatile?

So I came up with the idea of a “Painting” as a style for writing. A painting (with ink) has the property of being made gradually, one brush stroke after another. There is almost zero commitment required in each stroke, but all are intentional. In a painting, no one but the artist can say whether it is finished or not. In a painting, you can always repaint some areas to reflect new understandings. Also, no one can say “That painting is wrong”. The purpose of a painting is not to be right, but solely to express. In that sense, a painting is a live medium, free of boundaries and accessible to anyone.

So from this day, I’m structuring my writings as “Paintings”. That way I freed myself from the drama of correctness. I am free to just open any tool and just write anything, save it, come to it later, add something, remove another, etc. The message of the corpus is the corpus itself, the pure expression embedded into it, and not bound to the linearity of the ideas it conveys.

The issue of presentation (meaning that moment you expose your work to the public) can also be served by this style without changes to the work, with the simple disclaimer of the literary style involved. That way you let the public know how to digest it.

This approach was very liberating to me and opened my eyes to the fact that formats and practical tools (software for ex.) are not the only tools we have for expression. The styles in which we invoke our ideas are also tools, and tools that do not depend on practical interfaces or materials.

Think about all styles that exists: poems, poetry, tales, etc. Each serves a purpose. Each is a tool. Why not create more styles?

Here is a painting where I defined painting. It is similar to Free Writing, but I find a Painting to be even more liberating in essence, since it could also be made of non-sense if the author wishes. The “writing” aspect is only accidental and not a requisite, much like figures can emerge from abstract art.

Any corpus of text must always start stating its character.

Is this text a piece of evidence? Is it a report? Is that text poetry? Is that text a critic?

In what light are we supposed to read it?

I propose a new character for texts that will be open for anyone, from scientists to laymen, to use to express themselves freely.

I call this character ‘Paintings’. A text is painting with words much like artwork on canvas is a painting with ink.

In a painting, the artist must not worry to include anything that he is not concerned with.

The sole purpose of the painting it to be self-evident, self-contained, and not to be compared nor interpreted. It is what it is.

A painting as a text character is much the same. In a painting argument, the author (now an artist) is free to stroke any wisdom, unrelated, unsound as it may be, to convey his expression.

A painting text can also display any sort of sub-classifications. But further classifications are not a burden for the artist to carry.

Further classification is a particularity that is entirely a necessity for the viewer, or for society itself.

From the artist’s point of view, his painting is his expression and that is it. It begs no further context nor explanation.

No one looks at a painting and says “This is right.”, instead they might say “I like it.”. And further, people will be able to be inspired by some painting to do their own painting.

The only thing the artist is entitled to do is name its painting, just as I will name this painting as “The first painting.”

1 Like

The idea of blog-chains is an interesting form of publishing. It relates to the concept of piecemeal growth from Christopher Alexander also.

But in the context of your blog post I can’t help but wonder: isn’t this format increasing the possibility of theme-abandonment and instead of relieving the author of premature commitment to the published work, actually increasing the burden of commitment by spreading it through time? Doesn’t it devalue synthesis? How is the theme of a blog-chain any different from a tag?

In the light that blog-chains are a good way of achieving multi-author collaboration, I can see value in this format. But I hold the opinion that such a goal could be better served by the classic form of dialectics.

Say you come up with an affirmation like “X needs to change”. You then find a colleague who wishes to agree with you but holds different views than you do. You two can now, in a sequential series of informal arguments, work the way through agreement. It could be structured like a regular chat (left and right ballons) contained in a blog post updated as arguments are added. I see this way like an escape from the “idea-incest” that self-expression tends to assume when posed in series under a theme.

Overall, the issue of overcoming perfectionism that you mentioned in the post seems to not be fulfilled by the blog-chain format. I see it just as another way of perfectionism.

On the goal of “writing more” that you like to achieve, you have demonstrated that you do write constantly, so what “writing more” actually means is allowing yourself to publish more. In that sense, wouldn’t it be better if you actually perfected a form of writing for yourself that is actually acceptable for publishing? That way every note you take is a note that can be published. No matter how novel the format to which you transpose your thinking, transposing will always be the barrier that delays your public expression.

Sorry if such opinions seem intrusive, but I also wrote these comments directed to myself. The questions you are facing are common to me too.

There’s something to your “painting” analogy that resonates. Yes, part of it is just the approach to writing. What is the expected outcome.

Then there’s also something that doesn’t transfer for me: expectations towards quality.

There are still good and bad paintings. The fact that they can be deconstructed into individual strokes has more to do with the process than the style. I know it’s not what you mean, but you make it sound like we can all just pick up a paintbrush, draw a few random lines here and there, and end up with a Picasso. (Insert “only one line” joke here)

I do get your point about style though. I think I still see both Digital Gardens and Blog Chains as less of a different medium with a different style and more as intermediate results in a process which yields even higher standards of quality; what’s new is that these intermediate artifacts are also published and no longer just part of the writer’s process.

As I was publishing my blog chain articles to my blog and assigned tags, I was asking myself the exact same thing.

It’s definitely not obvious, but I do think there is a difference — less in the structure of these posts and more in the expectations communicated to readers.

Blog chains are somewhere on the spectrum towards “chaos streams” — the timelines and Twitter threads that explore micro-publishing and smaller feedback loops in writing. There’s a lot going on here, and I’m just beginning to understand how interesting and influential it all is.

You are absolutely right that I should say “publishing” instead of “writing” as I do write a lot and this is about sharing my thoughts with others.

Considering my writing earlier in the process as something publishable is good advice. Thanks. I was already headed in this direction with blog chains and that didn’t work out, so I guess I’ll try the gardening thing next… :slight_smile:

Hey all - I’m a new member. Great to be part of this community! Curious, are there any Tiddlywiki users here? I used tiddlywiki to create a small digital garden, but I’m thinking of trying another tool. Do people recommend Gatsby or something else?

1 Like

Just stumbled into Gardens which looks like a nice take on a slack+wiki+garden tool.

1 Like

I have experimented with a few garden environments based on a this Next.js + Mongo template on GitHub.

For Roam users here interested in building a digital garden based on it. Excited to share - a new service I’ve built that allows you to easily create beautiful static websites/digital gardens (like , , ) based on your Roam graph (takes just a few clicks :wink:)

1 Like