28 Nov 2014

In the talk “JavaScript luv Unicode”, @mathiasb really given a pretty good introduction to unicodes. This is a brief post on what I learned about uncodes from his talk. Please note that many of the following things are explaned in the context of JavaScript.

All the symbols in unicode standard will have a unique number called code point and a unique name. we don’t much care about the name in this post, but we use the word “code point” a lot. For example, the letter “A” is represented as U+0041 in unicode where the integer part 0041 is called code point.

A - U+0041
a - U+0061

The codepoint will span from 000000 to 10FFFF. Thats the whole range for possible unicode code points. In this there are reserved code points, code points which are not yet assigned and most used code points. For easiness, the whole unicodes are divided into 17 planes.

Basic Multilingual plane (BMP)

The group(plane) which consist of most used symbols are called Basic Multilingual Plane (BMP). This includes code points from 0000 to FFFF. All the language symbols, even malayalam letters are included in this plane.

Supplementary planes/ Astral planes

All the other 16 groups are collectively knows as supplementary planes. Astral plane span from 010000 to 10FFFF.

Hexa-decimal escape sequence

The symbols from 0000 to 00FF can represented in hexa-decimal sequences. It can have only 2 digits codepoints.

"\x41" // A  
"\x42" // B  
"\x41\x42" // AB  

Unicode escape sequence

The symbols from 0000 to FFFF can be represented in unicode escape sequence. It can have only 4 digit code point.

"\u0041" // A  
"\u0042" // B  
"\u0041\u0042" // AB

This can be used to represent whole symbols in BMP. So what about astral planes?

Right now the symbols in astral planes are represented as two unicode escape sequences which is called surrogate pairs. A pair of 16-bit code units. A surrogate pair will represent only 1 character.

"\uD83D\uDCA9" // surrogate pairs for unicode U+1F4A9 which represents PILE OF POO

In ES6, a new escape sequence will be intoduced to JavaScript to overcome this.

// only in ES6
"\u{1F4A9}" // PILE OF POO
"\u{0041}" // A  
"\u{0042}" // B  
"\u{0041}\u{0042}" // AB

Due to this representation of unicodes in astral planes, as surrogate pairs, there are lots of issues in JavaScript in handling uncode string manipulations, which will come in later blog post.

Finding Surrogate pairs

As per the unicode spec, there is formula to find surrogate pairs. The first part of a surrogate pair is called High surrogate and the second part is called Low surrogate.

high = Math.floor((codePoint - 0x10000) / 0x400) + 0xD800  
low = (codePoint - 0x10000) % 0x400 + 0xDC00  

Thus, you can find the high surrogate and low surrogate with the above formula. Now if you want to find the codepoint from the surrogate pairs, you can use the formula below.

codePoint = (high - 0xD800) * 0x400 + low - 0xDC00 + 0x10000
If you particularly enjoy my work, I appreciate donations given with Gittip.