minor fixes

Author: iliakan

Diff for: 1-js/02-first-steps/05-types/article.md

@@ -68,9 +68,20 @@ We'll see more about working with numbers in the chapter <info:number>.
 
 ## BigInt [#bigint-type]
 
-In JavaScript, the "number" type cannot safely represent integer values larger than <code>(2<sup>53</sup>-1)</code> (that's `9007199254740991`), or less than <code>-(2<sup>53</sup>-1)</code> for negatives. Technically "number" type can store larger integers (up to <code>1.7976931348623157 * 10<sup>308</sup></code>), but outside of the safe integer range <code>±(2<sup>53</sup>-1)</code> a lot of integer values can't be represented using this data type: some of them are "missed" due to the limitation caused by their internal binary representation. For example, all the odd integers greater than <code>(2<sup>53</sup>-1)</code> are "missed" in the "number" type.
+In JavaScript, the "number" type cannot safely represent integer values larger than <code>(2<sup>53</sup>-1)</code> (that's `9007199254740991`), or less than <code>-(2<sup>53</sup>-1)</code> for negatives.
 
-For most purposes <code>±(2<sup>53</sup>-1)</code> range is quite enough, but sometimes we need the entire range of really big integers without missing any of them, e.g. for cryptography or microsecond-precision timestamps.
+To be really precise, the "number" type can store larger integers (up to <code>1.7976931348623157 * 10<sup>308</sup></code>), but outside of the safe integer range <code>±(2<sup>53</sup>-1)</code> there'll be a precision error, because not all digits fit into the fixed 64-bit storage. So an "approximate" value may be stored.
+
+For example, these two numbers (right above the safe range) are the same:
+
+```js
+console.log(9007199254740991 + 1); // 9007199254740992
+console.log(9007199254740991 + 2); // 9007199254740992
+```
+
+So to say, all odd integers greater than <code>(2<sup>53</sup>-1)</code> can't be stored at all in the "number" type.
+
+For most purposes <code>±(2<sup>53</sup>-1)</code> range is quite enough, but sometimes we need the entire range of really big integers, e.g. for cryptography or microsecond-precision timestamps.
 
 `BigInt` type was recently added to the language to represent integers of arbitrary length.