Notes by Lex Toumbourou

Semantic Web was an extension to the World Wide Web, envisioned by Tim Berners-Lee, to make the Internet data machine-readable and understandable, allowing computers to process and "reason" about information in a more intelligent way.

It utilises a few interesting technologies like the RDF data model, the Linked Data set of best practices and Graph Databases, such as Triplestores.

Though many of the technologies are in use throughout the Web, for example, the use of JSON-LD to provide machine-readable metadata and Schema.org for SEO, it never really achieved the kind of widespread adoption hoped for partially due to the "chicken-and-egg adoption problem": content creators didn't have incentive to add semantic markup to applications without applications to consume it, and applications weren't created due to lack of semantic markup in use. Also, other problems have been identified, like "most web users were likely to provide either no metadata at all or else lots of misleading metadata meant to draw clicks", similar to the abuse of <meta> tags to fool early search engines.

I recall that people used to refer to it as Web 3.0 before the Web3 blockchain, which itself fizzled out, due to a lack of practical purposes beyond hype and scams.

The Semantic Web was also proposed as a form of Knowledge Representation, an approach to AI reasoning that has really been surpassed in favour of LLMs, which can reason (albeit imperfectly) over unstructured textual data.