Embedded Input Method

In order to answer the information department's requirements as well as to smash the monopolization of the foreign code in automobile in China, I make this research. Since the imputing way is based on smartphone roof, I absorb the liberty imputing and the frame of windows 2000SDK. However, the code standard and the code structure is secret, I make a research and design the code deeply and also I offer a complete code instance. On imputing the Chinese characters by keyboard, at the first beginning, we do it by characters and then by words. But we can know the developing direction is intelligence imputing, which is always the highlight in both academe and industry. From the analysis on the exerting arithmetic, intelligence imputing can be stat. and syntax; this code is based on the stat and also designed on considering the limit of EMS memory. Imputing characters by keyboard is used some changes of imputing code clusters and characters ISN clusters. The main factors that reflect the change between ISN and precision of Characters are the large number of the repeating codes. The purpose of the intelligence inputing is to choose one relationships from the excessive. So, the truth is we can input the code and then formed the inquired words. And later we choose the result according to the context, which is the probability stat take into the intelligence inputing. The above structure shows us the basic process of imputing characters. From the statistics, the related information between imputed characters and characters, words and words can be achieved by seeking the most the road. That is to say, we can regard a sentence as a process of seeking the superior roads. So the course that realizes the intelligence choosing is a seeking superior road one. Regarding to the words in code, we can static by training the word stock. And also we adjust them in the course of imputing. Something need be noted: with less word materials or too large room of parameter, some legal combinations often are neglect, although they have been calculated, which is very easy to lead to non-transformation among the context. In order to answer this question, we can adopt the way of flowing data. There are many flowing methods. While the linearity one, which is found by De.mori in 1990, is most popular for its expedience. This code can be realized according to the people daily and the user-defined stock, so the flowing questions can be resolved by it. In daily life, we set 1000 user-defined room for stock the user-defined words. As the using frequency of the words, some green words can be deleted freely, which can save much room and also can add new words. That can update the user words stock by alternating of people and mobile. The basis of intelligence imputing is keeping the fixed and common words. Such as 1000 characters and 1000 words can be fixed, others can be shielded. But they can be picked up, when users unlock all imputing. This can save much tome. But it also has shortcomings. If a user often use different words and often update the words, it will fail to show its effects. Another advantage is: it can fold all the characters and words into the memory to treat them. This can improve the inputing. The treatment is: all the data are stocked as 32 bite'unsigned int (Abbre: unit or UINT). It can be divided into 4 fields. Each field is relatively independent. While some information fells on next sub domains. E.G: the first one field is index table field. All the data is 32 bits'unsigned int (Abbre: uint or UINT). The initial 612 uint is for calculating. The first uint is for uint calculating in the entire whole developing uint array. the total 10 uint of the high 16bits, from 2nd to 11th, is the first uint index in the code: 0xxx,1xxx,2xxx…,9xxx. The lower 16bits is useless.