The start of figuring out what I should do for a final project
for A593

Mills declares the first part a reasonable start.

Unfortunately he's not provided any references, though I've
asked, so I've started reading chapter 4, Process Management
Subsystem, of:

 Goodheart, B. & Cox, J. (1994). The magic garden explained:
 the internals of UNIX System V Release 4, an open-systems
 design. Prentice Hall. 

which I happened to have lying around.


---------- Forwarded message ----------
Date: Tue, 26 Mar 2002 21:04:25 -0500 (EST)
From: cdent@burningchrome.com
To: jwmills@cs.indiana.edu
Subject: ideas for other projects


I set myself about the task of figuring out a project for myself.

Several ideas came to mind, related to things that I thought
might come up in class:

- various ways of doing stack manipulation
  - either something that puts a frame pointer on the stack for
    passing args and results on the stack
  - or self modifying code that tweaks the stack to manipulate
    flow control

- operating system issues
   - memory allocation
   - process management
     - multi-tasking
     - priority setting

To that end, how does something like this sound:

Take something, perhaps just the existing CVM code, modify it so
it gives up control at intervals, and build a simple process
manager that will run several calculators at different
priorities, managing memory and the stack.

And get as far as possible on that. If that goes well, create a
queue of incoming calculation byte streams, process them in
faux-parallel, garbage collecting as necessary, and picking up a
new byte stream off the queue as needed.

Today I don't have much of a clue how to do this, so I don't know
if this is feasible or too much. Does it seem reasonable?

What do you think?

Also, do you have any comments on those questions I sent about
class options for the fall?