Hello,

one group in the department I work for considers offering a practical
project group about online algorithms, and to spice it up by embedding
those algorithms in a general-use operating system. The question is:

Which of our kernel subsystems are easily replacable? ISTR reading 
about Disk i/o queueing and the scheduling algorithm - what about
page replacement?

The idea is not necessarily to produce code suitable for later inclusion
in NetBSD. The students would have to implement a suitable algorithm, or
several, and compare their performance and scaling behaviour (also with
the default algorithms of the OS).

Regards,
	-is