the milkmedicine webchain

webchain Specification

webchain is a distributed webring, where each member can nominate other websites, creating a walkable graph of trust.

In webchain, each member nominates others to the chain, who get their own nominations in turn. This creates a tree-like structure, where each node has a parent and potentially children.

The size of the webchain is not fixed, and is only limited by the half-life of the internet. A node may disappear from the webchain, taking with it all of its nominations. This is not a bug.

Each node has a responsibility to ensure that its children are trustworthy and don’t abuse the system. If a member or its dependents are found to be abusive, they can be removed from the webchain by members terminating their nomination higher up the chain.

Rules

  1. The tree starts with a seed, for example: https://mychain.org. There is one seed in the webchain.
  2. Each node in the tree represents a website, which is uniquely identified by its URL.
  3. Each node may define child nominations by including appropriate <link> tags in the <head> section of its HTML (see examples below).
  4. Each node can nominate up to n[1] other nodes, with any additional nominations being ignored. Nominations may be freely added or removed by the node owner, at their discretion.
  5. A nominated node can be any valid URL, including subdomains, paths, or even different domains altogether.
  6. Nodes may not nominate themselves or create cycles in the nomination path. If these cases are detected, the offending nominations are ignored.
  7. If a node does not respond in a timely manner, responds with a bad status code, or not at all, the node is considered not indexed. Non-indexed nodes are still a part of the data representation, but its subtree is not traversable until it comes online again.

Example node

<!DOCTYPE html>
<html lang="en">
	<head>
		<meta charset="UTF-8" />

		<!-- The webchain this node is part of -->
		<link rel="webchain" href="https://mychain.org" />
		<!-- Nominations made by this node -->
		<link rel="webchain-nomination" href="https://www.example.org" />
		<link rel="webchain-nomination" href="https://www.wikipedia.org" />
	</head>
	<body>
		example webchain document with two nominations
	</body>
</html>

In this example, this node has nominated two others: www.example.org and www.wikipedia.org. They are now part of the webchain mychain.org.

Note that simply adding webchain markup to your site does not automatically make you part of the webchain - the example must first be nominated by an existing member.

Joining a webchain

Let’s say you have a website, https://www.example.org, and you want to join the webchain at https://mychain.org.

  1. Find the webchain seed (e.g., https://mychain.org) and visit it to understand the community and its purpose
  2. Request nomination by contacting existing members with open slots, asking for a nomination.
  3. An existing member can admit you to the webchain by adding <link rel="webchain-nomination" href="https://www.example.org"> tag to their page.

The page is now part of the webchain. They may now nominate others by adding markup to their own page.

Crawling

Webchain is intended to be crawled from the seed node. The crawler should start from the seed URL, fetch the HTML content, and parse it to extract nominations.

Assuming the nominations are valid (not breaking any of the stated rules), the crawler should then recursively visit each nominated URL, fetch its content, and extract further nominations until all paths are exhausted. It may be useful to keep track of visited URLs to avoid cycles.

The crawler should also handle errors gracefully, such as timeouts or invalid responses, and skip over any nodes that cannot be reached.

It may be of interest to store metadata over time about each node, such as its join time, last seen time, and nomination history so that the evolution of the graph can be studied.

A note on visibility

The webchain isn’t very fun on its own. Visualization software will have to be created to traverse the graph and display it in a meaningful way. This includes handling of historical states, non-indexed nodes, and other metadata.

The webchain seed should be treated as a starting point for discovering the rest of the chain. A web page or application should be created to visualize the chain and allow users to navigate it.

Discoverability may be an issue, as visitors of any given node will have to know the webchain exists to browse it. Members should thus be encouraged to link back to the webchain seed node or other representations available.

If all parts work well together, a fun and dangerous social experiment will emerge.

Concerns

  • Since a nomination URL can be any valid URL, it is possible for a single actor to create a large number of nodes and nominate each other. Since the webchain is based on trust, such an attack is not likely to succeed unless the attacker can convince existing members to nominate them.

  • If a nominated domain expires and is purchased by a malicious actor, they inherit the trust relationship and position in the webchain without the original vetting process.

  • With each node having its number nominations limited, the webchain could become unable to grow if the edges don’t invite new nodes. This could be mitigated by allowing more nominations per node, but that would also increase the risk of abuse.

  • With the current design, members may only be in one webchain at a time. This is a deliberate choice to keep things simple. Allowing multiple webchains would increase markup and crawler complexity.

This is not an exhaustive list.

  1. The value of n is left unspecified. The webchain admin may choose a suitable value based on the desired growth rate and trust level of the chain. A common choice might be n = 3. ↩︎