table of contents (in Russian) 11 Декабря, №4




Compositional input on-line.
(brief description, details here)


    The idea of the compositional input is simple and obvious. Because in the faces of the Chinese characters there are repetitive elements, Chinese characters can be grouped according to the presence of certain elements, and then one can "ask" the computer to display characters depending on the presence of certain combinations of the above components. All. The rest - a programming tricks (i.e. a matter of the computer :). Let it work itself. In the local version it works for a long time (look here), but now You have the opportunity to "try it out for yourself" without leaving the network. Keep in mind that, of course, all this works obviously slower than would work locally on Your machine, and the functionality of the network version, of course, limited. But while the network version of one task is to show (and see :) what it is: a) possible, b) simple and c) is available.

    Generally compositional input was intended to solve 3 problems: 1) to facilitate the search of the Chinese characters, the pronunciation of which is not known, 2) to accelerate and simplify the input of the known characters, and (3) to simplify the process of remembering the Chinese characters. I think all of these three tasks compositional input solves successfully. Bowing to the people owning at least one of the classical methods of searching for Chinese characters ("Russian graphic", "the four corners", "key + strokes" ), before their hard work and perseverance, I dare to say that for a "not sophisticated Sinology" mind of the beginner that simple combinatorics components simpler and easier. And the ease and speed of typing and say no, if almost 83% of characters you can enter with two clicks of the mouse. What could be easier? And new-fangled touch device with a small screen, where the phonetic input is transformed into a "circus tricks"?..

    Incidentally, the relevance of the first task has already penetrated into "the consciousness of the people." Anyone interested can look at the MDBG website, where this year appeared search Chinese characters by components - "Show entries containing character components: (汉字部件)". Great work, even though the set of components, it seems to me, unnecessarily large (300 PCs.) and does not solve the second problem - the acceleration input, although this task is not standing in front of them.

    Structural hieroglyphics describes the rules "dismemberment" of the characters on components or decomposition rules, but in most cases the decomposition is obvious. Chinese character "字" consists of two components: "宀" and "子". The Chinese character "我" consists of components "手" and "戈", character "你" three components: "亻", "⺈" and "小". Let us call these components of Chinese characters  graphs. In this embodiment, the composite system has 200 units. Mind-blowing number to remember? But "classical radicals" no less - 214, and nothing, no one is crying :)... And the short table of graphs counts fewer, and only 3-4 of them are not "full copy" of radicals. More or less complete description of the tables of graphs and their variants can be found here .

So, each Chinese character is uniquely described by a linear sequence of graphs. But unfortunately, as a sin, the character "孽"("a sin"), for example, describes nine counts: 艹,丿,丨,乙,一,乙,一,辛,子. Fortunately, in the full description of the characters there is some "information redundancy", and one can uniquely identify the character by a shorter sequence of graphs. Let's call this shortened sequence of graphs that uniquely identifies the character marker. For the just-mentioned character "孽", consisting of 9 components, a marker is a sequence in only 2 count: 艹 and 丿. This is encouraging. In the entire set of 4341 characters which "knows" the composite input system, there is not one in which the marker would have been more than 3 counts. This means that using the search by markers, one can enter any character from a given set by three "clicks" of the mouse as a maximum. At that 72% of characters have markers from one or two graphs. Is it worth to remember markers? Especially when you consider that the first graph of marker is always the first graph of the full sequence of graphs of the Chinese character. If marker consist of 2 graphs, the second graph of marker is most often the second, or sometimes the last, graph of the full sequence of graphs for given character. And when marker consist of 3 graphs, this marker constitutes by the first, second and last graphs of Chinese character (with rare exception).

How to cope with all this things...

    As always, everything is not simple, but very simple.
    Click on the desired graph in the matrix. In the top window, you will receive the desired graph and in the next window, depending on the "Search Mode", appear "candidates for choice": Chinese characters that contain the desired graph. Also in a separate window indicates the number of characters that matches the selection criterion.
    Search mode:
    - "free", when the choice of characters is on the basis of availability of desired graphs in any combination in the description of the character;
    - "mark", when the selection is on the basis of strict compliance with the sequence of the selected graph in the marker.

    The link "Clear" removes all that can be cleared, returning the system to its initial state. Click on a Chinese character in second window, depending on the "Output Mode", gives You either the translation of the desired character (in a separate window), or throws the desired character into the right window, where you can "to pick up" the text by highlighting and copying it to the clipboard, and paste it wherever You want.

    The graphs table is divided into two parts for 100 graphs. The first half of the graphs table is displayed during the first page loading.  Link Shift replaces it in the second half, and the second replaces the first. After entering characters to the output window, or after the presentation vocabulary card, matrix is returned in the original condition. Graphs from the first half of the table, with respect to the frequency of occurrence, covers 90% of use, in other words, in nine cases out of ten You do not need the second half of the table graphs at all. Tested on Lenovo's Android. Rides. Why then the native browser not ship with site extremely desired font. But Chrome works all correctly. Sorry, I have no more devices for testing :(...

    That's all.

    Now you can return to the page of the On-line test and try to write something, for example, the first phrase from the training texts: 是 祖国 的 小 主人.
    Set Search Mode to mark, Output Mode to write, in turn, click on the graphs until you see the required characters in the window. Click on it and transfer it to the right window.
   
手,戈 - 我. 亻,门 - 们. 日,疋 - 是. 示,目 - 祖. 囗,玉 - 国. 白,勹 - 的. 小 - 小.丶,王 - 主. 人 - 人.Total: 9 characters and the end point, 24 mouse clicks.
    And this is still without any "tricks" such as typing whole words. And with such tricks for 10 characters, including the end point, 14 mouse clicks. Is it impressive? You can get  a writer
at the downloads and check.    
 
   
    Have questions? Welcome: support@poutko.ru


table of contents (in Russian)
Page begin 11 Декабря, №4