28 lines
1.2 KiB
Markdown
28 lines
1.2 KiB
Markdown
# Architecture
|
|
|
|
RETRO has a multilayer design.
|
|
|
|
At the heart of the system is a virtual machine called Nga. This emulates
|
|
a 32-bit stack processor with a MISC based instruction set.
|
|
|
|
The core RETRO language is stored as a memory image for Nga. The *image
|
|
file* contains this and is loaded on startup. It holds all of the compiled
|
|
words and data and interacts with Nga.
|
|
|
|
The third layer is the user interface. RETRO doesn't specify any required
|
|
I/O other than a console log capable of receiving a single character at a
|
|
time. Each host system can implement this and any additional desired I/O
|
|
by extending Nga.
|
|
|
|
# Specifics
|
|
|
|
On iOS the user interface is setup around an editor and an output area.
|
|
The editor extracts code from fenced regions, splits it into tokens
|
|
(elements separated by whitespace) and passes each of these into RETRO
|
|
for processing. After the tokens are processed, the console is updated.
|
|
|
|
The iOS code provides additional I/O functionality. In the `file:`
|
|
namespace there are words for creating and using files. There's a
|
|
`pb:` vocabulary for interacting with the clipboard. And a `clock:`
|
|
namespace which provides access to the system time and date. These
|
|
are mapped to higher opcodes outside of the core set used by Nga.
|