README

Fuzyll's Challenge of the Nao #2
********************************

So, the last challenge didn't go so well...  I only got 2 submissions,
and neither of them were fully complete.  One was able to actually
execute the "Hello World!" program, so at least the challenge wasn't a
total failure!  Regardless, I think it's probably time to move on.  I
had wanted to do a harder and more interesting challenge for #2, but
it kinda required everyone to understand how brainfuck works...  We'll
save it for later.

Your next challenge is to create an emulator.  The instruction set
architecture you will be emulating is big-endian and has 18-bit wide
instructions.  For now, assume that these 18-bit instructions are
padded to a 32-bit width to make it easier.  You should also assume
that each 18-bit instruction is made up of 4 discrete sections:

1. 6-bit opcode
2. 4-bit operand 1
3. 4-bit operand 2
4. 4-bit operand 3

Thus, the sequence 00000000000000100100101011110000 should be interpreted as:

- 14-bits of padding
- the 6-bit opcode "100100" (36)
- the 4-bit operand "1010" (10)
- the 4-bit operand "1111" (15)
- the 4-bit operand "0000" (0)

For now, all I want your emulator to do is interpret any binary stream
from a file (seriously, just pipe /dev/urandom into a file and read
from that) and display how it interpreted it.  So, given a file
containing the bytes above (actual bytes, not ASCII 0's and 1's), your
interpreter should print something like the following back:

OPC 36 | OP1 10 | OP2 15 | OP3 0

One line per instruction parsed.  What your interpreter does when it
encounters an "instruction" that contains "invalid" padding (anything
other than 14x 0 bits) I'm leaving undefined.  You may either print
"Illegal instruction" or just parse it anyway.  In the future, we'll
be removing the padding and making your emulator actually execute
18-bit instructions.

Submissions can be in any programming language.  I would prefer to get
links to submissions from GitHub (because you should all be using the
WCSC GitHub repo for your WCSC projects and, therefore, spend a lot of
time on GitHub anyway), but you may also just send me your code.  Your
submission should be able to properly interpret any input I give it.
As such, I only require your code as a submission and not whatever
streams of data you choose to test with.