In recent years, the model's proposal has suddenly, but what does this concept mean? From philosophical, our idea is the reflection of the external world structure in subjective consciousness. When we re-shoot back to the outside world, we get models about the external world. Therefore, in the broadest sense, the model is just a set of associations in our thinking. The problem is not what we need model or something is a model, what is not a model, and everything we realize is model, no matter how big it is with the real situation. What we can distinguish is what is a "good" model, what is a "bad" model. In object-oriented modeling, I often hear people say that XX is an object, YY is not an object. This is a wrong proposed method. All things are objects, they are symbols that we can manipulate in our mind. A paper is an object, when it is torn into a fragment, each debris is also an object. Because of the disturbance of random factors such as force, the way the letter is broken, and the final resulting debris is also random. We can say that the letter L is constructed by the three sub-objects of the debris A, B, and may also say that L is composed of debris D, E, F. The way the object is defined is endless.
We cognition the external world by comparison. When we face an unfamiliar concept, we always break down, and refactor it to truly understand it in the concept we are familiar. Maybe we can only understand what we already know! We have survived hundreds of millions of years on this planet, so that we can always compare those "new" structures with our "proven knowledge". Only when you face quantum theory and relativity, a person can truly realize how much you know, a innate poor. Cantor will play a comparative technology to the ultimate, and he is divided by comparison, but the number is infinite, and the result is crazy. Suppose the reality has a structure S, our model has a structural M. When S (through a certain simplification and topology transformation) can match M, we say that reality S is understood. More general, we look at the same thing from different angles or on different levels, thereby forming models M1, M2, M3, ... we can pass a logical pathway m1 -> m2 -> m3 -> S Understand S. The best model should construct the shortest path to the real structure S. A fresh example is the database. Under the von neumann architecture, the problem that the computer can directly answer: how much is stored in the 0x *** address (M1). In reality, the question we often need to answer is: Value is greater than ** records (M2). These issues are the value of the variable, not the address of the variable. The value of the database is being mapped by the index to provide mapping of M1-> M2. Early level database only provides an intermediate model H, the mapping of H-> M2 also requires programmers to complete, and it is ultimately inevitably eliminated. A model created an industry like this. The shortest path of the model map means that we'd better not create a new structure of "smart". In a well-designed system, we should not force "discover" or what to cognition. The design is not equal to creation, and one of the best designs should be an example of copying. It is best to make a change. Unfortunately, the reality will always have to change much. If you need to change, it is best to be locally changing, and we call it in mathematics. If the localization cannot be done when mapping, it often means that some essential difficulties have occurred. For example, the Table element in the HTML language. The form is essentially two-dimensional layout, and HTML text is a one-dimensional text stream, and there is a profound difference between them. When a table unit's ROWSPAN or COLSPAN attribute needs to be modified, it is only a local adjustment, which is reflected in the HTML code is a global correction. XML is never universal. A thing is clear, because we have a complete model M for it, so we can model the known things, because we can start from the model M, establish a similarity or even The same model can be. All difficulties are modeling for things we know less or nothing. According to Laplace philosophy, if we don't know what is the best choice, then all choices are equivalent. When using an iterative method, the usual approach is to throw a random initial solution, as long as the control strategy is correct, will eventually converge to the correct solution. Everything is not how "smart" the initial heuristic policy, randomness is the creative root, because only a chance can break the shackles of causal. A good initial solution is generally just a process of shortening the convergence, but does not affect the final convergence ending. A model is first provided is a term system or a vocabulary, a language. Let us capture those fleeting thoughts in a consistent way. Whether the model itself is correct is not necessarily important as it is imagined. Once the initial model is established, our awareness has a constant development and accumulation of the foundation, and it will eventually get the result of our initial expectations.
