What will Bi-Directional Links do to the Internet
Backlinks requires it's own Medium
Before we start talking, enjoy this pretty illustration by Maggie Appleton explaining uni-directional hyperlinks and bi-directional linking :
There is really no faster way to immediately appreciate the difference between Uni-Directional Hyperlinks and Bi-Directional links than Maggie’s illustrations.
Like Facebook built it’s social graph on the foundation of mutual acknowledgement between people, I want to explore making a content graph on the foundation of mutual acknowledgement between content. (At Metamind Co., we are directly building for this.)
For individuals and for some teams, bi-directional linking is realized in the form of modern graph-based knowledge management apps. We will be taking insights from these apps and applying to the global, social context of humanity.
Hyperlinks are mono-directional. When you hyperlink to a page, the destination page has no idea that you linked and cannot acknowledge the relationship you are trying to form. Hence, no user who discovers the other page first will be able to back-track to your page. This is sad! After all, your page does contain relevant content.
Backlinks, while a common data-point for recommended content and search results and a common feature of personal and team information management app these days, is absent for direct public maneuverability around the open web. People don’t have the agency to navigate the web through backlinks.
Surfacing backlinks to a page is a technically possibility, but it has yet to scale into a behaviour for any given environment, let alone the open web. We can take Twitter’s Retweeting as an example, but it is an untapped possibility on other popular platforms like YouTube for instance. And even in Twitter, it is not one of the common behaviors for people to follow-through on retweets, like how they engage with replies.
And yet to me, it almost seems obvious that there is uncaptured value in Backlinks to the order of what Hyperlinks are known to have. There are deeper relationships, novel business models and a whole new medium to be explored. Hyperlink creates the medium of Internet. By the end of this essay we will see how Backlinks leads to it's own Medium as well.
When inventing the World Wide Web, Tim-Berbers Lee did explore having (automatic) bi-directional linking on the open web. But it was never implemented because there were a ton of obvious and non-obvious challenges with it: whole new kind of spams and trolls, UX complications, too many backlinks, etc.
Too Many Backlink Requests
The first obvious problem with backlinks is if everything that links to your pages automatically becomes visible on your site, that’s a lot of uninvited dust settling in your online home. On the other hand, if this becomes permissioned and manual, that can be one-too-many requests for you to handle.
It might be feasible if you could filter out requests as per some reputational system and/or relationship system based on a growing social graph, but today, our social graphs stay siloed within social media platforms.
One can imagine today’s biggest social media platforms “connecting” to your website, and filtering bi-directional requests across people’s websites but they are hardly incentivized to build such an infrastructure. It almost seems like they don't even want you to know that a web outside their platforms exist. Twitter's algorithms demote tweets that contain links, Instagram eliminates links from it's posts altogether and YouTube wants you to auto-play to the next video in the very next breath, instead of pausing to check the video description full of useful links painstakingly put together by our favorite youtubers.
However, there is a new era coming. Recently, paragraph.xyz (a web3 substack-like platform) tapped into the social graph of Farcaster (a web3 decentralised social graph). Farcaster has a Twitter-like client called Warpcast running on their Protocol. What Paragraph did here is pipeline all conversations happening around paragraph.xyz-links when shared by the respective author in Warpcast.
As you can see in the image below, all Warpcast conversations are embedded right under the Paragraph posts. With similar accesses, we can speculate that in web3’s open, interoperable world, a social graph can determine whose content can backlink and becomes visible in your content spaces.
This is a good start.
It must be noted here that TikTok has made clear the flaws to the approach of approximating a content graph based on a social graph, and have dominated American social media for years now. Social Context is not irrelevant, ofcourse, but it cannot directly be the proxy for Content Graph.
Other metadata will be required. Perhaps, the kind of meta-work that people are doing to build their Content-Graphs inside Personal Knowledge Management Apps today.
In fact, even when some automated filters are deployed, I believe there is scope for new connections to be discovered and novel relationships via manual curating efforts. But what even is the incentive to curate Backlinks (and create a Content Graph), the way creators are motivated to do all the effort of injecting Hyperlinks?
Incentive for Curating Backlinks
First of all,
Writers do all the work to hyperlink parts of their content as it saves them from laboriously establishing the context, in order to make their ideas comprehensible to an audience larger than those who are very familiar.
Establishing context can often be an insurmountable task as there is always more context to a context and so on, which one usually discovers going down a rabbit hole of hyperlinks until you arrive at familiar grounds from which you can bounce back and progress with the content, or ofcourse, find new interestingness down the rabbit holes.
It increases the target audience of the content. When you publish something on the global stage of the Internet, it is the rabbit hole of hyperlinks that the reader follows to gain context when he has none.
Second of all,
Interestingness down the Rabbit Holes adds to the perceived value of a content to the user.
I can report from experience that the value of a piece of content increases when it helps one discover more valuable content. And creators know this. As VGR notes in his blog, the Rhetoric of the Hyperlink :
Real hyperlink artists know that paradoxically, the more people are tempted to click away from your content, the more they want to keep coming back. - Venkatesh G Rao
Discovery is indeed valuable. A lot of creators make themselves feel worth a monthly-subscription by meticulously serving curations in their newsletters. But, is discovery enough of an incentive to manually curate backlinks?
Building on Mutual Acknowledgement between Content
I think Discovery is a real, but barely scratches the surface for what can be done with Backlinks.
Accepting a backlink request is a moment of mutual acknowledgement between two parties. This moment can be seeded with contracts that forge valuable partnerships between their content.
The simple case is when if two paywalled-content mutually acknowledge each other, then readers of one can be given automatic access to another. This can be implemented with Token-gates.
Likewise, a bundling-contract across mutually graphed content can be bundled up and sold retail while involved parties drive the sale for each other. Consider this:
A free content F links to 3 paywalled content P1, P2, and P3. The moment a user likes F, he can be prompted to purchase the bundle of P1, P2, and P3 that pays back F too. The free content is essentially how the user discovers and is persuaded to buy the bundle. This way free content can get paid and paid content can get more reach.
Another possibility enabled by Web3 is that micro-transactions are now a reality.
Certainly, micro-transactions will be a powerful tool to eliminate spam requests and acting as a filter for only serious backlink requests.
To make a micro-payment while requesting backlinks may make a lot of sense for a mutualistic behaviour of the referenced author helping accurate usage of his ideas. Call it webbed-teaching.
With the speculative yet promising future of tokens the partnerships that can be forged between people via their content is a vast yet-to-explored space. A dynamic enabled by Backlinks can make such partnerships commonplace to the internet. However, if Hyperlinks creates the medium of Internet today, can we simply expect to add Backlinks and hope everybody will start using it. I don't think so.
Getting Backlinks to Internet runs into a big UX Problem
Hyperlinks has the intuitive UX of clicking-through based on a context supplied by the hyperlinked-text and the content around it. It is like choosing to open a box complete with a “notional” label outside.
But what is the natural narrative to click through a backlink? Do we expect users to also add supporting text along with curating backlinks? Without a persuasive, organic narrative to the UX of backlinks, it is no different than the “Related Content” section.
Hyperlinks gets you around the Web way more than links in the Related Content section. Hyperlinks makes linked-content a part of the ongoing exploration. Hyperlink is the medium of the web. On the other hand, one mostly consumes content from the "related section" as an afterthought to consume more.
Backlinks appear much later after the content is published, and the context to click-through has to be supplied then. In graph-based Personal Knowledgement Management apps like Roam Research, when you see backlinks under a Roam page (or under a Roam block), you can see the specific block from the page that links back. This block naturally supplies context for the backlink.
But content on the Internet doesn’t exist in blocks, yet. In the current state of our Internet, built around the document metaphor, splicing out such a “block of text” around a hyperlink to serve context for the corresponding backlink, not only seems technologically messy but also unnatural to the medium. People are not expressed in atomic blocks of Ideas.
Hyperlink is the native nature of Internet today, and Bi-Directional Linking is natural to the private little internet that Roam Research creates for a single user or a certain team. Writing in Roam, your thoughts get naturally segmented and committed to independently-legible blocks of text that can be discretely pulled in to supply context for backlinks.
You can't simply retrofit backlinks on an hypertext-linked internet. Our online expression will have to change, like it changes from writing in books to writing in blogs to writing in Twitter threads. Expressing in threads, in a string of threaded-nested blocks, seems like a medium within which backlinks can be had on the Internet.
This is precisely where we, at Metamind Co., are experimenting by enabling writers to publish their content-graph created in networked-tools for thoughts like Roam Research, LogSeq, etc. Inside these tools, the content is created in blocks and bi-directionally linked. With Metamind, creators will be able to publish these content graph and then we will evolve the platform/protocol to have bi-directional linking across content from different people. It is the shared Second Brain vision that many has given to this new Internet.
Subscribe if you like reading essays like this. Next up, I will be writing how Bi-Directional Linking bends Temporality. Talk soon.