Comment Sections

Comment sections are a staple of blogs and social media, for better and for worse. And they're not on this site, at least in it's current form. Lets talk about why.

Note

This garden page is written a lot more like a blog post than my usual writing. It is more of a reply to the discourse on comments that happened on the small web in 2025 and less evergreen than my other pages. I recognized while writing this that the way I referenced other pages came more from the perspective of introducing a new reader to my work. I don’t think any of this is a problem per se, but something I noticed and felt worth pointing out.

Comment sections on youtube, x, reddit, and so on all face some similar problems. Although the exact mechanics vary, these are all platforms that allow spending money to highlight messages, are designed for a small number of highly popular users/“influencers”, and generally don’t lend themselves to constructive and collaborative discourse. These comment sections are floods of strangers you’ll likely either never see again (non-influencers) or are unlikely to recognize you in the future (influencers), so there’s no reputation or social consequences to worry about. Indeed, I think the best places on any of these platforms are generally the smaller, more niche sub-communities within them.

The small web (of which this site is a part of) is decentralized, and thus whether or not a site has a comment section, or how that comment section is implemented, varies.

Many sites simply don't have comment sections, particularly static sites that lack a database to store the messages or account information in. Some may use a third party service to do that for them (such as the options listed here). Whether it's first or third party, you're typically still having to type in (and surrender) your information on each new site you wish to comment upon. The exception, of course, being systems that let you use an existing identity, such as your google or github account, to log in. But that just further increases the centralized power of those other companies in exchange for convenience.

The main “solution” to this problem is webmentions. These let you write a reply on your own site and notify the person you're replying to about your post. You can keep your version up even if the recipient of the mention doesn't decide to display your reply below their post. They can't revoke your reply, as you have full ownership over your data. You are not beholden to any individual or company continuing to exist or remain in good standing.

Unfortunately, and as others have pointed out, webmentions are not very accessible. You have to have a website, for one, with either a backend or a third party dependency. That website will need to include microformats2 markup, which is also non-trivial to setup. There are some services that allow for anonymous webmentions, but those then also give up the data sovereignty benefits. Suffice to say, it's a lot for someone just wanting to comment on the article they read.

IndieAuth is also an alternative that would allow you to log in using your own website as your identity, but that still requires the recipient to have a backend, and the replier to own a website.

All of this points to comments on the small web not having a current good solution that maintains data sovereignty while still being accessible. Sure there are compromises to make it more convenient at the cost of data sovereignty, but I think we can and should have a solution that is easy and accessible while still being secure and that lets you own your own data. In fact, this should apply not just to comments but all communication online - be it short, long, tied to a certain context, or fully independent musings.

Unfortunately, whether its a missing comment section or one you don’t wish you to use, this makes blogs feel a bit closed off. There was discourse about having comment sections or whether its okay to just point people towards making their own blog posts if they want to respond publicly (even for short “I liked this post” thoughts) or reaching out over email if they want to respond privately. I appreciate both stances, and note how frequently people added the caveat to the effect of “you can choose to have comments or not on your own blog, you do you”. But notably, replies are always an option: they might just have to live on your own blog. I think from there, you get back to the argument that requiring a blog (or depending on a 3rd party) shouldn’t be required to participate in discourse. I’d like to see a way for people to own their replies without depending on 3rd parties, while preserving the right of writers to choose whether to feature those replies on their posts.

This site is a digital garden; a collection of pages that link wiki-style and are not just “post and forget”. They get returned to and updated, pruned, reorganized, and otherwise transformed over time.

The way I have this garden structured, a reply to an article wouldn't properly belong. The reply would be a whole page, and quickly replies would overwhelm the map of pages. Yet, putting the replies on just a single page would also be difficult, as they often would bridge between multiple topics and anyhow would become divorced from their context as a reply, which I find useful to retain.

If I change the page, separate things out into different pages, or any number of other things, the comment will also become stale and divorced from its context.

I think we can find a better solution than comment sections that would better integrate into the network of a digital garden, particularly compared to having a bunch of comments with only one association, that being to their parent post.

Better communication online

Comment sections, as well as the content they are replying to, make up communication. I think framing this whole discussion on the merits of comments and how to improve comment sections perhaps disguises the bigger point of improving communication online in whole. Indeed, all the concerns raised about making comments also apply to the content itself. I think trying to solve the problem of communicating online - constructively rather than adversarially, lets us achieve all the seemingly incompatible goals and restraints discussed thus far.

Accessible discourse

The first requirement for our alternative should be maximizing accessibility. My litmus test for this is someone should be able to write and reply to other people's writing while retaining full ownership over their data, without spending any money nor trusting a company to not shut down nor enshittify.

This may sound like an impossibility, but it's not. I've described such a solution called the Agentic Fediverse, but really all you need is a way to identify yourself and your content in a way that doesn't depend on a service continuing to exist indefinitely, of which there are many solutions that exist today. From there you just need a way to deliver your content to others, but once you control your identity and data, the delivery method doesn't matter, and can change with new technologies, protocols, or resources. It might rely on a freely hosted relay server today, but use a P2P network like meshtastic tomorrow.

Everything is a reply

Comments are replies, but often so are the articles themselves. This may be in a literal sense, in the form of ”re:“ articles, but also in the references contained within the article. It's building off ideas that already exist, as a citation or to rebuke or form any other kind of association.

Our current structure of articles, garden post, replies, comments, etc. don't really encapsulate all these associations as robustly as they could. I suggest reading the Chat Glue comic for a vision of what might be able to construct this ever growing web of associations. It describes the concept in relation to a conversation, but it turns the conversation into a living document, and I think this living document can grow into a whole forest of documents that associate with each other and is built for collaboratively constructing new knowledge in a way our current social media (blogs included) simply can't.

To be clear, these associations are more than just the comments and response posts. A reply could be a suggested edit, a reaction, tagging a post, a “highlight”, or anything else. It can be directed at a portion of the document rather than the whole. Hell, perhaps they could even be displayed after the paragraph it's “replying” to, rather than at the end (for short replies that are highly contextual, perhaps). In theory, even people not interested in writing should still be able to form associations between texts via these “replies”, and contribute to the connectedness and organization of the network.

Reducing spam, hate, and marketers

There are some comments worth discouraging, like those listed. For its faults, the barriers to entry afforded by indie web comment sections, they are much nicer to read and far easier to moderate compared to other social media. But I’d like to achieve the same effect while preserving that accessibility.

Part of removing the reliance on centralized services does mean leaving centralized moderation behind as well. There is nothing person A can do to remove content by person B, no matter how justified it would be. This is a problem, although not a new one. The indie web is already decentralized, and a person already cannot prevent another person from replying to their article on their own website. The best person A can do is moderate their own page, and not show any comments or webmentions from person B. And, to a lesser extent, this applies to centralized social media as well, where at best you can get the platform in question to remove content or even ban person B, but you still can’t ban person B from posting on other platforms. I’ve given my own perspective on decentralized moderation and think there are still things we can do to empower individuals to not see the content they do not wish to see, but limiting the spread of bad actors is ultimately more of a culture problem.

For something like this, where I would want to create an environment built to foster community and consensus building and preventing hate, I think the biggest thing is to design a system where it’s hard to go viral. The small web is largely positive and constructive because of its size - when you’re in a small community, you recognize names and have an actual reputation and social consequences for being toxic, in a way you just don’t have on a platform like X where you can essentially be an anonymous troll “drive-by posting” and never have it impact you. I’m not suggesting we wrap everyone in an echo chamber, but that shrinking the web a bit (while notably still encouraging diversity) can help you actually get to know people, feel for them, and be nice.

Funnily enough, a lot of decentralized moderation is about adding barriers that make it harder for spam or other bad-faith posts to spread. It’s worth differentiating the earlier proposals that remove barriers in the name of accessibility from proposals to add barriers in the name of reducing spam and hate. These barriers shouldn’t be based on requiring people to meet certain standards in order to participate - e.g. you shouldn’t require a website or money - but it’s okay to rely on mechanisms of trust building, reputation, or social consequences (in my opinion), even though they may limit the reach an individual might have (after all, that’s sort of the point). That said, I’m aware that “trust building, reputation, and social consequences” can all be affected by the biases of the people involved and, intentionally or not, exclude people based on their race, gender, or neurotype (I’m autistic myself, and am keenly aware of how social consequences can be abused in a neuronormative culture).

In practice, this means designing a system where sharing ideas (e.g. political movements) is easy but sharing specific posts or people is hard. I believe the best way to do this is removing algorithmic feeds, ensuring at the very least that the content we see is written by or as a response to someone we’ve gone out of our way to connect to. I think there’s still considerations to explore here, and I’m not going to pretend I already know the perfect solution to a theoretical “anti-virality” social media, but I’ve written my own ideas in what I call Digital Locality.

Orchard

I've been workshopping these ideas for awhile, and orchard is my vision for what this all looks like. Now that you've read all this, you already understand the motivations behind the project, which is what a lot of that page is about. But I recommend checking out the sections at the bottom referring to prototypes. And if you're interested in the technical details, the rest of the page is good too.

That said, it’s not a perfect solution and I consider it a long-term project that will require various prototypes/experiments to work out the details of how all of this should work. Early on, for example, it might not be very accessible for people without an orchard client on their device. Ideally I’d love for something like a browser extension or app that can allow you to send replies/comments on articles you find online that get stored on device and sent to a relay server, while requiring minimal setup for the user in advance. It’d also be nice to have some sort of system where existing small web sites can become orchard documents, with the site having an “identity” that others can verify wrote those documents, and which can later vouch for the identity of the actual author’s account, if/when they start using orchard.

I’ll close off this by saying I actually have pretty high hopes for the ability of orchard. And since part of the aim is to have a smaller web with more consistent connections, it becoming super popular isn’t necessarily a requirement for it to be successful - it’d just need to be meaningful and useful to those who do use it. And with that goal in mind, I think orchard has a lot of potential to allow people to learn together collaboratively and constructively, form relationships, and create a large network of knowledge that could help discoverability and organization of even niche topics.