I already considered that, but until now i did not implement many peripheral devices. But when implemented those, it should be possible. But at the moment i'm rather thinking of an VHDL netlist emulator in TeX.
In the 1990s, when I was in college, I wrote a 6502 emulator in a text editor macro language (the text editor was QEdit, a DOS editor that was popular at the time). It was completely useless but a lot of fun.
There was a very low hard limit on the number of jumps (32, if I recall correctly) that complicated things. I eventually had to resort to a two step process, first translating 6502 code to a simpler virtual machine (using and abusing the find-and-replace command), and then executing that simpler code.
I wonder if this could be modified to control actual hardware? It looks the `\write18` command can be used to execute shell commands if LaTeX is run with the `--shell-escape` option. So in theory you could intercept all writes to the GPIO control registers and use a shell command to send bytes over the serial port to a physical MCU (or something like the SparkFun SPI Shortcut.)
Imagine a LaTeX document that, when processed, reads a series of measurements from a sensor over the serial port, then renders nice graphs directly to a PDF! :)
I think this is the minority view of LaTeX: it's built on top of and is a subset of TeX, not a superset. No one would ever consider \def to be LaTeX for example.
The dataref package is intended to be useful (and it acutally is). Avremu was only a long trip in sweden without internet, an editor and a reference manual.
sometimes, cool is enough :) I might actually want this though. C is much more convenient than TeX to code in, and yet sometimes you have no choice. Until now.
http://dmitry.gr/index.php?r=05.Projects&proj=07.+Linux+on+8...
Emulate that system instead, and now you can theoretically boot a Linux system inside of TeX.