Better ReadMe

ReadMe.md

@@ -9,23 +9,29 @@ console.log(yid())
 // -> 1517049989798-7496988299172
 ```
 
-A `yid`:
+A `yid` is:
 
 * always 27 chars long
 * has two parts:
   * a timestamp
   * a random string
 * is of the form `\d{13}-\d{13}`
 * starts off with `Date.now()`
-* uses `Math.random()` for the second part
+* uses a substring of https://www.npmjs.com/package/math-random for the second part
 
 ## Why? ##
 
 Why another ID generating library?
 
 I already wrote [zid](https://www.npmjs.com/package/zid) and [flake](https://www.npmjs.com/package/flake) (a long time
-ago) and they all have uses. The use for this one is to generate FAST but UNIQUE IDs. A secondary property is that they
-are approximately sortable across servers.
+ago) and they all have uses. The use for this one is to generate FAST but UNIQUE distributed IDs with no central server
+to talk to and no co-ordination required.
+
+A secondary property is that they are approximately sortable across servers.
+
+I got the idea from Google Keep, since the notes have IDs as follows, whereby the first secion is just `Date.now()`:
+
+* '1517050593526.6629835825', e.g. https://keep.google.com/u/1/#NOTE/1517050593526.6629835825
 
 ## Author ##