BCI Bytecode 
*************

.. contents::

Assembly and Bytecode
=====================

Unlike machine code (and other bytecode) BCI bytecode has
dynamic opcodes. This means that bytecode is **not**
necessarily portable. 

This makes sense since the BCI instruction set can be
extended for applications.

If one wants to share code that should run on any BCI it
should be shared as assembly. The assembler will then use
the local interpreter definition and generate suiting
bytecode.

The Dynamic Instruction Set
===========================

The BCI comes with a set of prepared instructions. These are
complete and provide a way to do basic operations like
routines, loops and branching.

The methods are organized in a binary tree internally. To
build the tree in a comfortable way there is an autoinserter
that can insert up to ``1023`` methods into the tree.

The autoinserter creates the opcode basing on the order of
the method that he inserts.

Byte Code Interpreter Definition
================================

A Bytecode Interpreter Definition consists of two mayor
parts: The memory definition that defines the number of data
registers (up to 63), the number of memory words (up to
65535) and the number of program memory words (up to 65535).

The second part defines the commands. The definition
contains bot the order of the commands (see `The Dynamic
Instruction Set`_) and the required arguments.