Fractal Writing System for Unicode

- Graphic symbols currently in use in natural writing systems can not be read by computers. Reading for natural writing was on the scanners tested, but failed. If the glyphs are printed, then error rate of read is lower, but if handwritten, the error rate is very high and are therefore useless.

- But we may use glyphs constructed according to computer science. In this way the error rate will be close to zero (and at computers and at people). This even if handwriting is negligent.

- For these glyphs have been established numeric codes (by the standard Unicode). For this ground we will call "Unicode glyphs”. Using these codes we will can find Unicode-glyphs for the entire world. See Figure .

Development of Unicode glyphs using fractals

- We will explain how the pattern is constructed. The pattern will develop fractals. See Figure . The model consists of a fixed segment AB. At the end (B) of the segment (AB) we will can draw two segments (BC and BD) perpendicular to AB. If the segment BC is drawn, then it shows that the most significant bit (from Unicode glyphs) is one. If he is not drawn, then this shows that the most significant bit of the glyphs is zero. Similarly, the segment of BD indicate the value for least significant bit.

Figure 1 = Graphic characters easy to write / read by computers and by the,h3> people

- With this pattern we will can represent only 2^2 = 4 glyphs. They are drawn to a 1 landmark. Now, we will do the first stage of development of fractals. We will represent a grammar with (2^2)^2 = 2^4 = 16 glyphs. We will repeat this pattern at the edges C and D from pattern. We will repeat this pattern at the edges C and D from pattern 1. It will be rotated through 90° and simultaneously it will be reduced in proportion to 1 / 2. Thus we will find those 16 glyphs Unicode from the first stage (see 2 landmark).
These glyphs are also insufficient. We will develop the second stage of these fractals (see 3 landmark). We will repeat this pattern at the four edges of the fractal. We will add the pattern rotated and again reduced to 1 / 2 (in total reduction is of 1 / 4). Now we obtained (2^4)^2 = 2^8 = 256 Unicode glyphs. These are sufficient to represent the glyphs of Latin grammar. They are sandartizate in ASCII of long time.

- However and these 256 of Unicode glyphs are not sufficient to represent all the glyphs existing in all languages. From this we infer that necessary the third stage of development of these fractals (see benchmark number 4). We will rotate the pattern and again we will reduce it at 1 / 2. Now the pattern is reduced in total at 1 / 8. We obtain (2^8)^2 = 2^16 = 65 536 Unicode glyphs. They are sufficient.

- If, however, will require more than 65 536 glyphs, then we can proceed to the fourth stage of development of these fractals (see marker 5). Now, we obtain (2^16)^2 = 2^32 = 4,294,957,296 of glyphs. The fractals method allows their development to infinity.

Remark 1: Thickness of the lines is reduced at every stage of development of fractals. However the creators of fonts must to can to change these thicknesses. This does not affect their reading. We will use a uniform thickness. Arial font do so.

Remark 2: We have a pair of successive bits. They are zero. Both are at the end of a segment for junctions. Therefore, that segment becomes useless. We will remove them. This glyphs will become more visually recognizable.

Unicode glyphs for latin alphabets

- Latin glyphs from Unicode are match those in the extended ASCII code. Code Unicode expressed on glyphs using two-digit numbers under the 256 base for numbers, so are the four-digit in hexadecimal (under the 16 base of numbers). The first two hex digits of the Unicode Latin glyphs are zero. They are expressed only by two digits in base 16, because we do not express the first zero digits.

Figure 2 = Examples of latin glyphs and chineses glyphs

- Latin writing system has less as 256 glyphs. Therefore we can use fractals to second stage of development (benchmark 3). We will explain the arrangement of bits in the glyph. We represented the arrangement for these bits in a glyph which has the code FF(in base 16) = 255(in base 10) = 1111 1111(in base 2) (see Figure 2 landmark 6).

- As example (see the landmarker 7); we will choose the phrase "Berlin is Germany's capital.". Under each Unicode glyphs we have written its binary code. Above the glyphs we wrote its the font New Roman. Above the glyphs New Roman is written its adequate code in hexadecimal. The binary code for glyph each shows (according to rule stated above) us which segments should be drawn and which not.

- At the landmark 8 is extracted only the respective phrase (without codes, so without numbers).

- At the landmark 9 we replaced the glyph (which expressing separation between words given by the ASCII code) with glyph which has the null code (0000 0000(in base2) = 00 00(in base 16) = 0(in base 10)). In this case the lines of structure becomes unnecessary. We will find again the white space for separator. In this mode the text is more readable, because this separator is the same as in the usual writing systems.

- At landmark 10 is written this phrase with Unicode glyphs for the lowest possible dimension. This the minimum size is explained in the landmark 16. It has 11 pixels height and width of 5 pixels.

- At the landmarks 11 and 12 is the same sentence, but we used large Unicode glyphs (18 × 10 and 22 ×10 pixels). We will read them easier.

- At the landmark 13 is illustrated the same sentence, but it is handwritten. With Unicode glyphs handwritten we can write sentences in a single line (without we lift pen from paper). This increases the speed of writing. However handwriting in a single line is NOT mandatory.

General Unicode glyphs

- With these we codify numerically the all glyphs the current language. They are very many and therefore we must use now the third stage of fractalisation (see the landmarker 4). Their codes have the first two nonzero hexadecimal digits. We will have to express them in glyphs. See marker 14. These four-digit hexadecimal give 16 binary digits (16 bits). At the fractals we deduced that they should be arranged as in the landmark 14 (at beginning from left to right and then from top to bottom).

- As example; we chose some Chinese glyphs. The 0C 4E glyph can be found in the Unicode table at line 4E 0C and at column 00 0C (therefore 4E 00 + 00 0C = 4E 0C). The glyph 4E 17 is at 4E 00 line and at 07 column. The chinese glyph 4E 27 is drawn at the line 4E 20 and at the column 07 (4E 20 + 00 07 = 4E 27). We transform these codes from hexadecimal to binary (see the benchmark 15). Then where the bit = 1 we will draw a segment. Where the bit = 0 we will let empty. Thus we obtain the respectiv glyphs Unicode chinese as on figure. First it is written in the smallest size (13 × 9). After the equal sign is written bigger (with size 26 × 18).

- And the general Unicode glyphs are very easy to handwrite and they are easily recognized (and without errors) by humans and computers. Both the computer and we can read the glyphs without to count theirs size.



- Please write to address: arghirescu_alexandru@yahoo.com