diff --git a/docs/user_defineable_self_namecall_type.md b/docs/user_defineable_self_namecall_type.md
new file mode 100644
index 0000000..99f752d
--- /dev/null
+++ b/docs/user_defineable_self_namecall_type.md
@@ -0,0 +1,169 @@
+# User Defineable ``self`` Namecall Type
+
+## Summary
+
+
+The aim is to allow simple definition of ``self`` when defining functions with ``:``
+
+```lua
+function Class:TestMethod()
+	self.@1
+end
+```
+
+
+## Motivation
+
+**Autocompletion is the motivation**.
+
+See the "Alternatives" section to understand the motivation as well.
+
+The point is to be able to define ``self`` in **namecall** defined functions as well.
+
+
+
+
+
+
+## Design
+Yet to be found.
+
+
+## Drawbacks
+
+Wouldn't see one for now.
+
+## Alternatives
+
+### Best Alternative
+
+```lua
+local Class = {}
+Class.__index = Class
+
+
+function Class.new()
+	local obj = setmetatable({}, Class)
+	
+	obj.Value1 = 1
+	obj.Value2 = 2
+	
+	return obj
+end
+
+type ClassObject = typeof(Class.new())
+
+
+function Class:TestMethod()
+	self = self :: ClassObject
+	
+	-- Now "self" has Value1 and Value2
+end
+```
+
+This ``self = self :: ClassObject`` is the easiest way that I know about. Based on my analysis. The default compile optimization mode will not count that part into compilation.
+
+Which is good, which means that this current way doesn't have any impact in anything. **But it is repetitive**.
+
+However, one will be able to index ``Class:@1`` to **differentiate** methods defined by ``:``.
+
+
+
+### Second Alternative
+```lua
+local Class = {}
+Class.__index = Class
+
+
+function Class.new()
+	local obj = setmetatable({}, Class)
+	
+	obj.Value1 = 1
+	obj.Value2 = 2
+	
+	return obj
+end
+
+type ClassObject = typeof(Class.new())
+
+
+function Class.TestMethod(self: ClassObject)
+	-- "self" has Value1 and Value2
+end
+```
+
+**This is still repetitive**
+
+The bigger issue is that this is not autocompletion friendly.
+
+```lua
+local Test = Class.new()
+
+Test:@1
+```
+This will show it in the autocomplete. But that's not the issue.
+
+The issue is when you use the pure Class table and index it with ``:``
+
+```lua
+local Class = {}
+Class.__index = Class
+
+
+
+function Class.new()
+	local obj = setmetatable({}, Class)
+	
+	obj.Value1 = 1
+	obj.Value2 = 2
+	
+	return obj
+end
+
+type ClassObject = typeof(Class.new())
+
+function Class.TestMethod(self: (typeof(Class) & typeof(ClassObject)))
+	-- "self" has Value1 and Value2
+end
+
+Class:@1
+```
+
+Notice how ``ClassObject`` is not resolved at that time.
+
+If one wanted to do ``Class:@1`` to filter out all namecall methods in their code, without creating a constructor. They simply can't.
+
+But this is a **metatable** in a class-construction context. Even Lua designs these OOP functions like that https://www.lua.org/pil/16.2.html
+
+
+
+### Strict Type Alternative
+This is an alternative that I captured off from **react-lua**.
+
+```lua
+local Class = {} :: {
+  -- This is the thing that defines self
+	[string]: (self: string) -> any
+}
+Class.__index = Class
+
+
+
+function Class.new()
+	local obj = setmetatable({}, Class)
+	
+	obj.Value1 = 1
+	obj.Value2 = 2
+	
+	return obj
+end
+
+type ClassObject = typeof(Class.new())
+
+
+function Class:TestMethod()
+	self.@1
+end
+```
+
+``self`` is now a ``string``. By defining ``[string]`` the automatic type resolver won't automatically put the actual things defined in the code into that table as well. That's why it isn't that good, but it's very smart.