docs: update README, link to docs.

README.md

@@ -2,22 +2,22 @@
   <img src="docs/v8serialize_logo_auto.svg" width="300" alt="The v8serialize logo. Monochome. Large &quot;V8&quot; and smaller &quot;serialize&quot; in a handwritten style, with the 8 stylized to look like a snake.">
 </p>
 
-# `v8serialize`
+<p align="center"><em>Imagine having <a href="https://developer.mozilla.org/en-US/docs/Web/API/Window/postMessage"><code>postMessage()</code></a> between JavaScript and Python.</em></p>
 
-Read & write JavaScript values in [V8 serialization format] with Python.
+<table align="center">
+  <tr>
+    <td><pre lang="shell">pip install v8serialize</pre></td>
+  </tr>
+</table>
 
-[V8 serialization format]:
-  https://chromium.googlesource.com/v8/v8/+/refs/heads/main/src/objects/value-serializer.cc
+<p align="center"><a href="https://h4l.github.io/v8serialize/en/latest/">Documentation</a></p>
 
-## Byte order/endianness
+---
 
-V8 uses the native byte order when serialising data. This library explicitly
-uses little endian. This is because:
+# `v8serialize`
 
-- The vast majority of systems using V8 are little endian
-- Because the serialized byte order is native, when people use it to store
-  persistent data they are probably assuming little-endian systems will read it
-  later.
-- I don't have a big-endian system or VM to test against
+A Python library to read & write JavaScript values in [V8 serialization format]
+with Python.
 
-In principle there's no thing to stop prevent adding big endian support though.
+[V8 serialization format]:
+  https://h4l.github.io/v8serialize/en/latest/explanation/v8_serialization_format.html