First, a bit of history.
Web 1.0 is a network for accessing content posted on websites by their owners. Static HTML pages, read-only access to information, and the primary joy are hyperlinks leading to pages on this and other sites. The typical website format is an information resource. The era of transferring offline content online: digitizing books, scanning images (digital cameras were still rare).
Web 2.0 is a social network connecting people. Users, fully immersed in the online space, create content directly on web pages. Interactive dynamic websites, content tagging, web syndication, mash-up technology, AJAX, web services. Information resources give way to social networks, blog hosting, and wikis. The era of online content generation.
It is clear that the term 'web 1.0' emerged only after the arrival of 'web 2.0' to designate the previous internet. Almost immediately, discussions began about the future version 3.0. There were several visions of this future, all of which, of course, were related to overcoming the shortcomings and limitations of web 2.0.
Netscape.com’s Jason Calacanis was primarily concerned with the low quality of user-generated content, suggesting that the future of the internet lies with 'gifted individuals' who will begin to 'create high-quality content' (Web 3.0, 'official' definition, 2007). This is a reasonable thought, but he did not explain how and where they would do this, on which sites - certainly not on Facebook.
The author of the term 'web 2.0', Tim O'Reilly, reasonably suggested that a person is not a necessary intermediary to host information online. Technical devices can supply data to the internet. Moreover, these same technical devices can read data directly from web storage. Essentially, Tim O'Reilly proposed to associate web 3.0 with the already familiar term 'Internet of Things'.
One of the founders of the World Wide Web, Tim Berners-Lee, envisioned the future version of the internet as a realization of his long-held dream (from 1998) of the semantic web. His interpretation of the term won out — until recently, most people mentioning 'web 3.0' meant precisely the semantic web, that is, a network where the content of web pages would be meaningful to computers and machine-readable. Somewhere around 2010-2012, there was a lot of talk about ontologization, and countless semantic projects emerged, but the result is well-known to all — we are still using the internet version 2.0. The only concepts that truly survived are the Schema.org semantic markup and the knowledge graphs of internet giants like Google, Microsoft, Facebook, LinkedIn.
The failure of the semantic web was masked by new powerful waves of digital innovation. Public and media interest shifted towards big data, the Internet of Things, deep learning, drones, augmented reality, and, of course, blockchain. While the first items on the list are predominantly offline technologies, blockchain is fundamentally a network project. At the peak of its popularity in 2017-2018, it even claimed to be the new internet (a notion repeatedly expressed by one of the Ethereum founders, Joseph Lubin).
But time passed, and the word 'blockchain' became associated not with a breakthrough into the future, but rather with unfulfilled hopes. Naturally, the idea of rebranding arose: let's not talk about blockchain as a self-sufficient project, but instead include it in a stack of technologies that represent all that is new and bright. Immediately, a name was found for this 'new' concept (though not new), 'web 3.0'. To justify this lack of novelty, it was necessary to also include the semantic web in the 'bright' stack.
Currently, the focus is not on blockchain, but rather on the infrastructure of the decentralized internet known as web 3.0, which consists of several core technologies: blockchain, machine learning, the semantic web, and the Internet of Things. Many texts that have emerged over the past year regarding this new incarnation of web 3.0 detail each of its components, yet there is a fundamental lack of answers to essential questions: how do these technologies come together to form a cohesive whole, what do neural networks have to do with the Internet of Things, and how does blockchain relate to the semantic web? Most teams continue to pursue blockchain (perhaps in the hope of creating a cryptocurrency that can rival Bitcoin, or simply working off investments), but under the new banner of “web 3.0.” Nonetheless, there are at least some discussions about the future rather than baseless hopes.
However, not all is bleak. Here, I will attempt to briefly answer the questions posed above.
Why does the semantic web need blockchain? Here, it is not about blockchain itself (the chain of cryptographically linked blocks), but rather the technology that ensures user identification, consensus validation, and content protection based on cryptographic methods within a peer-to-peer network. In this context, the semantic graph as such a network receives a reliable decentralized storage system with cryptographic identification of records and users. This is not just semantic markup of pages on a free hosting service.
Why does a hypothetical blockchain need semantics? Ontology is fundamentally about organizing content by subject areas and levels. This means that applying the semantic web onto a peer-to-peer network—simply put, organizing network data into a unified semantic graph—enables natural clustering of the network, which allows for horizontal scaling. The layered organization of the graph facilitates the parallel processing of semantically independent data. This is data architecture, not just indiscriminately throwing everything into blocks and storing it across all nodes.
Why does the Internet of Things need semantics and blockchain? With blockchain, it seems straightforward — it serves as a reliable storage solution with a built-in identification system for actors (including IoT sensors) based on cryptographic keys. On one hand, semantics allows for the segregation of data streams into subject clusters, effectively relieving nodes. On the other hand, it makes the data sent by IoT devices meaningful, thus independent of applications. One can forget about the need to request application API documentation.
And now we need to determine the mutual benefit of merging machine learning and semantic networks. It's quite simple. Where else but in a semantic graph can we find such a colossal array of validated, structured, semantically defined data in a unified format, which is essential for training neural networks? Conversely, what analyzes the graph better than a neural network to identify useful or harmful anomalies, for example, to isolate new concepts, synonyms, or spam?
And this is the kind of web 3.0 we need. Jason Calacanis will say: I told you it would be a tool for creating high-quality content by gifted people. Tim Berners-Lee will be pleased: semantics rules. And Tim O'Reilly will also be right: web 3.0 is about 'the interaction of the internet with the physical world,' about erasing the boundary between online and offline, when we will forget the phrase 'log into the network.'
My previous approaches to the topic
- (2012)
- (video, 2013)
- (2015)
- (2019)
Source: habr.com
