knyte knyte knyte space space space
EN RU

Semantic Web

Семантическая сеть

is a synthesis of knowledge management and computer science. In this article we give its detailed determination.

Methodology

In general, it is hardly possible to measure knowledge numerically. But if we consider knowledge in the context of its specific digital representation, we can identify some basic data structures that can be used to describe knowledge. Their structure and size may well be determined.

Knit

Similar to the bit for measuring information, the knit is the smallest unit of knowledge for a digital device. According to the postulates of knowledge engineering, a knowledge base can be represented as a set of semantic entities. According to the practices of computer science, each entity can be designated through a globally unique identifier. This identifier can then be used as a data structure to describe the knit. The size of the identifier determines the size of the knit. Let us use uuid v4 for identification, then the knit values will look like this:

ac2fb456-5936-48dc-934c-1bfcf6bd71e6
8a048d4c-93c0-45b8-8a36-e8a3cd80164d
71076d29-d4ff-430a-b9bd-9c9ee6e5584e

Thus, the size of knit in binary representation = 16 bytes = 128 bits.

Modern uuid standards like uuid v7 are also suitable for knits.

Knyte

Similar to the byte for measuring information, the knyte is the smallest unit of knowledge available to human perception. If byte is the number of bits needed to encode one character in the text, then knyte is the number of knits needed to fully describe one semantic entity.

node: entity id → knit → uuid v4
initial link: entity id → knit → uuid v4
terminal link: entity id → knit → uuid v4
content: content id → knit → uuid v4

Thus, the size of knyte in binary representation = 4 knits = 64 bytes = 512 bits.

Content Map

is a collection of pairs content id - textual value.

sample content map

ac2fb456-5936-48dc-934c-1bfcf6bd71e6: "first entity"
8a048d4c-93c0-45b8-8a36-e8a3cd80164d: "100500"
71076d29-d4ff-430a-b9bd-9c9ee6e5584e: "link"

Knoxel

Similar to pixel in computer graphics, knoxel is a way to display knyte within a semantic web interface.

sample knoxel

Semantic Web - legacy definition

Starting from 1980s scientists and researchers tryed to find ways to implement Semantic Web. One of most popular approaches belongs to Tim Berners-Lee. Back in 2007 he proposed migration from World Wide Web (also invented by him) to Global Semantic Web (also known as Giant Global Graph or GGG in short). Implementation of GGG was supposed to use a bunch of HTML-like languages:

As we can see now, the idea has not been widely adopted.

Semantic Web - modern definition

Using definitions of knit, knyte and knoxel, we can define the semantic web as a collection of interconnected knytes.

sample semantic web

From knowledge engineering, it is known that human knowledge can be represented as a semantic web - this is the knowledge level, that is, the highest abstraction known to us. Then, the semantic web can be represented as the graph shown above - this is the information level, that is, a form accessible for interpretation through sensory organs or computer algorithms. And finally, the semantic graph can be stored in a database, using a well-defined amount of memory - this is the data level, where storage, copying, and transmission of information is implemented.

Let's calculate how many bytes are needed to store the graph shown above. It consists of 13 knytes and 8 content units containing text with a size of (1*4 + 2*8 + 3*12 + 2*13) = 82 bytes. In total, storing the graph will require (13*64 + 8*16 + 82) = 1042 bytes.

Thus, through this example, we managed to connect the most abstract level of knowledge with the most concrete level of data. In perspective, this gives us the opportunity to optimize digital communications and implement reasoning artificial intelligence based on existing computer systems for data processing.

Homogeneous Recursive Semantic Graph

Semantic web visualisation shown above is called homogeneous recursive semantic graph. As classic math graph is designed to visualize node-edge structures from maths, homogeneous recursive semantic graph is designed to visualize semantic web. It's homogeneous because consists of one type of elements - knoxels. Knoxel can by node, edge or node and edge at the same time. It's recursive because links can link not only nodes, but another links as well. And it's semantic because every element of the graph represents some knyte, also known as semantic entity.

Metalanguage

It's well known in knowledge engineering that any natural language can be represented as a semantic graph. It's well known in computer sceince that any programming language can be represented as a semantic graph. If we unite these semantic graphs into one, we will get Metalanguage.

Metalanguage - is a single language for people of all nationalities and computers of all architectures.

Sentence once written in Metalanguage (metasentence) can be automatically translated to any natural language or to machine code for computer of any architecture. Metabooks, metasites and metaprogramms may be assembled from metasentences.

Classic communication in modern society

Meta communication in near future

Metalanguage able to replace all APIs, natural and programming languages in the field of digital communication.

Examples of Semantic Web