Sharing linked thinking on the web: methods for helping readers follow along

Publishing your thoughts on services like Obsidian Publish affords new kinds of opportunities for sharing linked thinking1 on the web. It seems to be easier (or at least more intuitive to do) than publishing conventional articles. This kind of writing also facilitates the accretion of knowledge in a way that breaking thoughts down into standalone articles seems to discourage.

Yet, there are some challenges. These ideas all work really well in terms of input, but I don’t think it’s so straightforward when it comes to output. Readers typically expect some method of easily keeping up with your contributions. Perhaps we need a paradigm shift from conventional publishing models to linked-thinking publishing, but until that happens, I think most readerships will be confused when they try to follow a writer along on a knowledge base.

That is, unless the writer provides some kind of help.

Far as I can tell,2 authors of linked thinking sites can do a couple of things to help readers follow progress. Here’s what I’ve come up with. (These options aren’t necessarily exclusive of one another—you can probably use multiple. They are also not ordered in any deliberate way.)

1. Updating an Index page.

One way to emulate “blogging” via linked thinking sites might be to keep an Index page. Every time you post a new item, link it there, perhaps with a date or some other kind of ordered identifier. That index becomes your homepage. Readers can check in to see when things are new.

2. Hybrid Blog–linked thinking site hybrid.

Publish a linked-thinking site. Maintain a blog within the site, perhaps by using a Blot-like service to watch a specific folder for new blog posts.

3. Posting a changelog.

Perhaps using a blog–linked thinking site hybrid or simply with “Changelog” pages in the knowledge base, track and annotate the changes you’re making in a specific report and share those reports as often as is appropriate.

4. Externalize the update mechanism.

Don’t provide readers with help tracking changes in-site. Push updates via other forms of social media—tweet out “Just added a new section to my page on abductive reasoning!” (or trumpet it on Mastodon or whatever) instead.

5. Embed RSS functionality.

I don’t think this is easy, but technically savvy folks may be able to set up an RSS service such that new (or modified) pages get pushed out via RSS. (Perhaps this is a good feature request for Obsidian Publish…?)

What are folks doing to help readers follow changes in this newish format? What have you seen? What do you think is the best model?


1: These sites used to be called “wikis,” but I think what we’re seeing people publish as “mind gardens” no longer fits the definition of wiki—which, to me, is a single dedicated repository of information about a given subject. A wiki should basically be all referential fact, while linked thinking sites seem like more than that! So, somewhat following Nick’s work, I like the idea of a “linked thinking site.”
2: I don’t actually have a linked thinking site—yet. I am still thinking about what such a thing should look like (obviously).

16 Likes

This has been on my mind lately. I like your thoughts on this topic.

I have had a traditional blog for years and have always struggled to write content for it. Maybe it always felt too formal. I’d get writer’s block all the time and would rarely finish a post. It just wasn’t fun.

I have started writing notes in Obsidian, and the process has become fun. For a while, I was using Blot, a great service for transforming markdown into blog posts. But then Obsidian Publish was introduced, and I immediately jumped on board.

Now that I’m using Obsidian Publish, I’m struggling with the challenges you mentioned.
Here are some things I’m trying.

  • Using the home page to link to the main category pages. (MOC pages).
  • First level pages are like MOCs. They contain a map of various subpages.
  • The actual content pages are still similar to a typical blog post page. In my experience, most people coming to my site would never even see the home page. They would land a specific page. I feel like this is where linked pages will shine. In a typical blog post, a visitor would read your article, and if they’re interested enough, they may click to your home page or browse some categories for other content. I feel like a well-linked page could potentially be more engaging. Especially with “Andy Mode” enabled where the pages stack. But will visitors find that type of layout confusing? I may ask some friends to check out my site and see if it is not very clear, to get their perspective.
  • Once I get things going, I want to try posting updates on Twitter, as you suggested in #4.

I don’t have much content yet. I’ve posted a few blog posts and have a few notes I hope to polish up and publish this weekend.

2 Likes

Thanks for these thoughts Ryan! It’s definitely something I have been thinking through (and concerned about) while watching this sudden explosion in PKM stuff.

If I am reading and understanding your points here, it seems to me that you are essentially envisioning a linked data PKM schema and a linked data enabled (semantic) blog.

If you’re not as familiar with linked data, it is essentially web 3.0 as Tim Berners-Lee argued. I won’t go into all the nitty-gritty of linked data (it can get really complicated very quickly) but there are a good amount of teaching tools out there (including my presetation here slides 16-24 here).

But I think a good way of thinking about it is thinking about the Recipe schema, which is marked-up formatting that allows a user to plug in a URL to a cooking site and save their own copy (lots of sites allow this now).

Another example is the Google Knowledge graphs, or the way that google has been linking to specific hilighted passages of text, like this for example :

https://github.com/mbakeranalecta/sam#:~:text=Semantic Authoring Markdown brings the,text editor -- like markdown.

I point to that repository because it contains a draft specification for semantic markdown. I would love to see something like that happen. The ability to do something like [[Barack Obama]] in your own PKM and have a knowledge graph or certain facts about Obama rendered in edit view = killer feature.

A interoperable PKM schema that was accepted across the web (or by a majority of players) would allow posts on different sites to be collated together easily, and cleanly and without a need for formatting. There are some tools to render tweets to markdown and then pull them together, for example.

This would allow for a variety of different platforms, technologies, and so on to be put together.

2 Likes

Thanks to you both for the replies!

I think learning from feedback is gonna be key. If the writer is used to linked thinking, it will be hard to see how someone new to the format receives it.

@brimwats The semantic web is somewhat related, but I don’t think it has explicit solutions for this problem, does it? (I work in conceptual modelling and data crowdsourcing, so I play in fields adjacent to web 3.0 stuff.)

The root problem here is that readers’ mental model of interacting with a writer is books, blogs, and columns—periodical formats. You subscribe, and then you receive new content when it comes out. That doesn’t quite work with a linked thinking model without some of the mechanisms I described at the top of this thread.

It is a great thought, though. Whether linked thinking == the semantic web might be an interesting debate, but I think it’s outside of scope. Still, the shape of the model of both things are the same: a graph of information. If you update a web 3.0 “site” to add new semantic data to it, how do the site’s users learn about the update?

2 Likes

The root problem here is that readers’ mental model of interacting with a writer is books, blogs, and columns—periodical formats. You subscribe, and then you receive new content when it comes out. That doesn’t quite work with a linked thinking model without some of the mechanisms I described at the top of this thread.

Thanks for the clarification! Maybe a model that would be useful is the doke.li which does allow subscription to blocks/pages/etc (IIRC!)? Your comparison here to periodical literature is illustrative and I’ll take a look at some of the LIS literature on how libraries handle changes as well!

2 Likes

dokie.li is fascinating. And yes, share anything that comes to mind!

1 Like

I’m looking to migrate my blog DAObase.org from Ghost to Obsidian Publish and have been thinking about this too. You’ll notice from the DAObase.org/discover page that I’ve been wanting to enable concept-oriented exploration of the blog for a while now.

These are my current thoughts for translating an Obsidian wiki into a newsletter:

  1. Host the vault in GitHub and use the commit history as a changelog

  2. Deliver the changelog to readers via one (or all) of the following methods:

  • Block Referenced Digest: Write a standard newsletter format, summarizing the changes that were made since the last digest, block referencing to the changes.
  • Copy Above Digest into Substack: After publishing the above digest to Obsidian Publish, copy paste into Substack. I can confirm that all of the publish.obsidian.md wikilinks are preserved when pasting into the Substack composer, which means the newsletter itself is a funnel into your Obsidian site, and you benefit from Substack’s discoverability. I haven’t tested how effectively block references/embeds paste.
  • RSS Feed / Automated Newsletter: GitHub automatically generates an Atom feed of your repo’s commits. This could be used in combination with the buttondown.email API, which can “programmatically generate emails”. I’m not sure how effective an automatically generated list of commits would be vs a manually curated digest, but I’m thinking I could offer both as separate mailing lists. I’m also curious if I could somehow tag specific commits to help curate which commits make it into the automated update.
4 Likes

Thanks for the detailed writeup. I think you’re onto something—rolling up committed changes and then reporting them in itemized fashion seems interesting.

I think you also might be able to use Releases (a GitHub feature) to package or “version” updates. Readers who want to follow voraciously can see every commit, and then readers who just want the big (summarized) changes can see releases.

2 Likes

Hey all, this is a really neat conversation! I’ve been mulling over some of these ideas and it came up in discord also (Discord). I thought I would re-post my thoughts here, where it seems relevant:

"Probably my favorite thing about andy’s notes (About these notes) are the conversations that reference his notes around the internet. I didn’t spend much time in his notes at first, but then people started posting links to various topics around the internet. However, this can break when the publisher makes changes. (e.g. Say Bob sends Alice a link to a note that no longer or exists or makes a reference to a point in the note that has changed)

We’ve seen this with APIs, which now often contain version numbers in urls. I wonder if something like that could be added to publish to help preserve the expected context of the link (but perhaps instead of version numbers they are the commit hash IDs…or have both and publishers can treat note publishing like software releases; e.g. new versions may fundamentally change things, while smaller releases do not).

Bonus: If the publisher does not like a previous version, specific notes at specific commits can be deleted or privatized."

[[ https://notes.short2things.io/versioning-published-notes/ | context ]]

4 Likes

An interesting conversation. I came across it as a new Obsidian user who likes the idea of publishing notes but who also has an existing blog. I’ve always equated blog with RSS for the benefits it provides, and was looking for ways to see how people handle it with Obsidian publish.

This article Does Your Blog Really Need to Provide an RSS Feed Anymore? generated some useful ideas. While obviously just their numbers, my experience is that most people don’t know what RSS is. Generating a periodic email with a change log is what most can handle. Sadly, if I were a user, I’d have to do more clicks to get to content, nevermind having to write summaries and an email as an author.

At present I’m trying to get my Obsidian Publish site hooked in under my existing blog site. That’s a mix of Gatsby, Gatsby Cloud, Netlify and of course Obsidian. I’ve come to learn the Gatsby Cloud link intermediary is redundant but I’m still testing. With this model I’d have blog articles which automatically generate RSS, that would reference a growing body of supported notes. I may then create an email model on top of that.

Another thought that came up re RSS is that it is not the means by which I find new content. That’s always Google/DuckDuckGo search. From there, if I see something I like I’ll look for RSS.

2 Likes

Obsidian now displays a date created and last modified feature. It seems like there’s space for a plugin (or perhaps a core feature) that could help solve some of this problem by wrapping up some links for which notes have changed in the last (specified) time period as well as which notes are new. Even as the notes author, I think I’d find that list quite useful.

4 Likes

Following up from this, are you seeing your obsidian pages showing up in Google searches? I am published on a sub-domain but nothing is showing which is a bit concerning. I have a blog but don’t want to create a link for every single page.

1 Like

I haven’t started using it yet so can’t test or confirm, though I also find that concerning. I thought I had read somewhere that SEO wouldn’t be affected by using Obsidian

It could be that I don’t fully understand it yet but I’d expect links from my blog to get into the Obsidian brain and for those links to be followed. It could be that it’s just all too new (< 2 weeks) but I am seeing new blog posts get picked up.

Possibly also that the redirect from my site to publish.obsidian.md is causing things to not be linked.

I don’t have the ability to get the insight to work out what’s really going on.

Hey @dcb if you are on the Discord server, it might be possible that Licat can help you figure out what’s going on. There shouldn’t be any issues with Google picking up publish sites that we are aware of.

2 Likes

Thanks. I’ll give it a few more days to make sure it’s all settled down.

Maybe it’s a bit of Baader-Meinhof since I saw the RSS help thread yesterday, but I ran into this reddit post for MkDocs Newsletter. I think the interesting parts for you are:

Every time you build the site, the plugin will inspect the git history and create the new newsletter articles under the docs/newsletter directory and configure the Newsletter section.

The entrypoints for the authors are:

And this one, about only adding certain types of commits:

The plugin assumes that you’re using the Angular semantic versioning format to create the commits. Adapted to a documentation repository such as the digital gardens, the structure would be:

{type_of_change}({file_changed}): {short_description}

{full_description}

[…]
For example:

feat(adr): introduce the Architecture Decision Records

[ADR](https://github.com/joelparkerhenderson/architecture_decision_record) are
short text documents that captures an important architectural decision made
along with its context and consequences.

Only changes of type feat, fix, perf or refactor will be added to the newsletter. The reader is not interested in the others.

Maybe there is enough ideas there to borrow from the existing code (that only works with MKdocs sites I assume) to make it work for any git-based blog. It might also be a thing to consider for a Publish feature request: Adding a small text box to the modal where you select the files you want to publish, and parse its contents according to some format to generate RSS feeds similarly to the post above.

1 Like

Hi @rscottjones, where are the date created and modified that you mention? I can’t see it anywhere on my Obsidian Publish page.

I only see it on the desktop version.

The date metadate is not embed in your notes, it’s associated with the marddown file and can be seen in any file explorer