Internal data representation and design problem. Koller's operations. An example of designing a kettle using a set of Koller. Criteria.

Internal data presentation and design problem

Having answered these questions, we will come closer to the answer, how to build a system that manipulates knowledge. We'll have to start with the answers to the simplest questions. So far, there is no ready answer to these questions, but the last five years have given new approaches and methods for solving these problems. After the turbulent 60s, and the lull of the 80s, optimism reappears.

How does knowledge exist within us? In the form of what? What is the structure of knowledge? Without fixing knowledge, we can not use them.

1) Thought form. This is not a text, not a formula, not a picture. It's all together. And not just stacked in a pile, but connected to each other. This is a multifaceted structure that can manifest itself by any party. Every time one that is convenient for use. And there is a dynamic in it. And there are connections with other thought forms. And these connections appear and disappear. Thought form is an element of knowledge. Every thought form is a well distinguishable whole.

2) The thought form is alive. She is changing. It represents the properties of an object that imitates from the real world. Or abstraction, if this thoughtform stands on top of the endless pyramid of other thoughtforms.

3) Thought forms are many, they are added, detailed (hierarchically), removed, assembled into structures to solve the problem, mapped, then disintegrated.

4) The medium stores thoughtforms, operates them, binds them into systems and hierarchically, details them, interprets them. The purpose of the environment - the solution of problems, tasks. For this, the medium selects the necessary thought forms, connects them. They form a new composition with new properties, and the environment interprets this combination of properties. The result of the interpretation is the solution of the problem.

Imagine a thought form as a complex model (formula, picture, text, link, sound, and so on). The model has the same properties as the thought form. It is dynamic, potentially contains in itself all the information about an object, imitating it, communicates with others, transforms, details on models connected again with each other, and enlarges. The model can be solved. The model has no input and output. They appear during the binding when solving a specific problem. The model may consist of new models related to each other. The model may be associated with other models of relationships. A model with links can be declared a new model.

All this description is reminiscent of system design. Something consists of something. Gathers and understands. It becomes complicated, communicates in new systems. Somewhere there is a constructor. And someone from the designer collects something on the working field. At the bottom - the simplest objects. By binding, they form larger ones. Then abstract objects appear.

Modeling helps to revitalize the engineered system. Interprets each element according to the rules of operation. Together they show new properties. During the simulation, the answer appears, the result of the problem.

Thus, for a start, we note that, at a minimum, thinking resembles designing objects and modeling projects in our mind.

A good design system is a prototype of artificial intelligence as a first approximation. And what is the "good"? And this means that the abstract (rather than narrow subject), object (rather than functional) interpreter, contains the possibilities of operating with objects (and manipulations, and descriptions), has all possible editors.

Therefore, further it makes sense to talk about design systems. There are two extremes:

1) A very large number of items. From this huge number you can do anything you want, but there is a high level of work. For example, from atoms you can do everything, but hardly anyone will decide on this.

2) A small number of items. This gives a low complexity in assembling the system, but it is very difficult to assemble other systems - the elements have a low degree of universality. The car "Ford" can be divided into 10 parts, but then no matter how you collect, "Lada" will not get.

How to find a middle ground? There should be a small but universal set of elements. But how to find it? There are more than ten systems offering various universal kits.

Koller's operations

More about Koller's operations

An example of designing a kettle using a set of Koller

Criteria

Requirements for the criteria:
Connection with time (the only non-renewable resource).
Measurability (quantitative): scale, rating.
Comparability (a single unit of measurement, comparability in time, values ​​should be relative).
Must assume the possibility of development, reflect it.
Independence from each other, their number should be minimal.

Criteria classification:
1) Functional (show how well the function is performed)

productivity (processing speed, physico-chemical parameters of the processing intensity);
accuracy (measurement, hit, processing);
reliability (reliability, durability, persistence);

2) Technological (saving labor)

the complexity of manufacturing;
technological capabilities;
use of materials;
the ability to break the product into elements;

3) Economic

consumption of materials;
power consumption;
information support costs;

4) Anthropological

ergonomics (full use of human capabilities);
beauty;
security;
environmental friendliness.