I expressed the idea of the inevitability of a global network transition from a site-centric structure to a user-centric one back in 2012 ( Philosophy of evolution and the evolution of the Internet, or in abbreviated form WEB 3.0. From site-centrism to user-centrism ). This year I tried to develop the theme of the new Internet in the text of WEB 3.0 - the second approach to the projectile . Now I am posting the second part of the WEB 3.0 article or life without sites (I advise you to look before reading this page).
So what is it? There is Internet in web 3.0, but no sites? And then what is there?
There is data organized into a global semantic graph: everything is connected with everything, everything follows from something, everything is specific, noticed, changed, created. The last two points about “follows” and “someone” remind us that the graph should not be object, but subject-event. But there will be a separate story about this (for a preliminary see the Subjective-Eventual Approach ). Now it’s enough for us to understand that the semantic graph web 3.0 is not a static set of knowledge, but a temporal one, fixing the relations of objects and actors of any activity in their temporal sequence.
Also, speaking about the data layer, it should be added that the global graph is necessarily divided into two unequal parts: a tree of models describing the connectedness of actions, concepts and their properties (corresponding to the set of terminological axioms of TBox in OWL), and a subject graph containing concrete fixation events values of the properties of things and actions (a set of statements about ABox individuals in OWL). And between these two parts of the graph an unambiguous relationship is established: data on individuals - that is, specific things, actions, actors - can be formed and written into the graph only and exclusively according to the corresponding models. Well, and, as already mentioned, the global graph - first of all, its model part and, accordingly, the subject - is naturally divided into segments according to thematic areas.
And now from semantics, from data, we can move on to discussing the second epithet of web 3.0 - “decentralized”, that is, to describing the network. And it is obvious that the same semantics should dictate the structure of the network and its protocols. First of all, since the user is the generator and consumer of the content, it is natural that he, or rather, his device, should be a network node. So, web 3.0 is a peer-to-peer network whose nodes are user devices.
To save in the data graph, for example, a description of a certain individual, the user must form a network transaction according to the existing model of the concept. Data is stored on the user's device and on the nodes of other users subscribed to this model. Thus, exchanging transactions on a fixed set of models on which their joint activities are implemented, participants in this activity form a more or less autonomous cluster. It turns out that the entire global semantic graph is distributed distributed across subject clusters and decentralized within clusters. Each node, working with various models, can belong to several clusters.
Describing the network level, it is necessary to say a few words about consensus, that is, about the principles of data validation and synchronization on different nodes, without which a decentralized network cannot work. Obviously, these principles should not be the same for all clusters and all data, because transactions in the network can be both legally significant, and service, trash. Therefore, the network implements several levels of consensus algorithms, the choice of the necessary is determined by the transaction model.
It remains to say a few words about the user interface, about the semantic browser. Its functions are trivial: (1) graph navigation (by thematic clusters), (2) search and display of data by domain models, (3) creation, editing of data and sending network transactions by relevant models, (4) writing and execution of dynamic action models, and, of course, (5) storage of fragments of the graph. This is a short description of the semantic browser functions and is the answer to the question: where are the sites? The only place the user “visits” in the web 3.0 network is his semantic browser, which is a tool for displaying and creating any content, any data, including models. The user himself determines the boundaries and form of display of his network world, the depth of penetration into the semantic graph.
This is understandable, but still where are the websites? Where to go, what address to type in this same “semantic browser” to get to Facebook? How to find a company website? Where to buy a T-shirt or watch a video channel? Let's try to figure out specific examples.
Why do we need Facebook or another social network? Obviously, for communication: to tell something about yourself and read, see what others are posting, exchange comments. At the same time, it’s essential that we don’t write to everyone and don’t read everything - communication is always limited to tens, hundreds, well, several thousand virtual friends. What is necessary for organizing such communication within the framework of the described decentralized network configuration? That's right: create a community cluster with a set of standard action models (record, send a message, comment, like, etc.), configure access rights to models and invite other users to subscribe to this set. Here is the facebook. Only not global Facebook, dictating the conditions to everyone and everything, but a customizable local social network, which is at the full disposal of the cluster members. The user sends a transaction to the network according to one of the community models, say, his comment, cluster members subscribed to this model receive the comment text and write it to their repositories (attach to the fragment of the subject graph) and display it in their semantic browsers. That is, we have a decentralized social network (cluster) for communication between a group of users, all of whose data is stored on the devices of the users themselves. Can this data be visible to users outside the cluster? This is a question about access settings. If there is permission, then the content of community members can be read by a software agent and displayed in the browser by anyone who searches the graph. It should also be noted that the number and complexity of cluster models is not limited by anything - everyone can customize the community to meet the needs of any activity. Well, it’s obvious that users can be members of an arbitrary number of clusters, both as active participants and simply by subscribing to individual read-only models.
Now let's answer the question: how do we find the company’s website? The answer is trivial: the place where comprehensive data on all companies is located is the corresponding sector of the semantic graph. Getting to this place will help browser navigation or search by company name. Then it all depends on the user - what models for displaying the data he needs: a brief presentation, full information, a list of services, a list of vacancies or a form for reporting. That is, the company for its presentation in the semantic graph should use a set of standard models for sending transactions to the network, and then the data about it will be available for search and display. If it is necessary to customize and expand the company's presentation on the network, it is possible to create your own models, including design ones. There are no restrictions, except for one thing: new models must be built into a single tree to ensure the coherence of data in the subject graph.
A trivial solution for e-commerce. Each product (mobile phone, T-shirt) has a unique identifier, and the manufacturer enters the product data into the network. Naturally, he does this only once, signing the data with his private key. A company that is ready to sell this product places in the semantic column several statements made according to the standard model about the price and delivery conditions. Further, each user independently solves the search problem for himself: whether he searches for the right one among the goods that a seller known to him can provide, or compares products of the same type from different manufacturers and only then selects a convenient supplier. That is, again, the place where the selection and purchase of goods takes place is the semantic browser of the user, and not some website of the manufacturer or seller. Although, of course, both the manufacturer and the seller have the opportunity to create their own models for displaying goods that the buyer can use. If he wants, if he finds it convenient. And so, he can do everything with the help of standard models for searching and displaying data.
It is worth saying a few words about advertising, about its place in the semantic network. And its placements remain traditional: either directly in the content (say, in videos), or in models for displaying content. Only between advertisers and owners of content or models is the middleman eliminated as the owner of the site.
So, the scheme of functioning of the semantic decentralized network, presented from the position of the user, is extremely unified: (1) all content is placed in a single global semantic graph, (2) the recording, search and display of content is carried out according to concept models that provide semantic data connectivity, ( 3) the user’s activity is implemented according to dynamic models, (4) the only place where the activity occurs is the user's semantic browser.