No overall "Project Plan"
  -- BUT more detailed Investigative Report
     including approach section & list of projects in prio order
	(by what solves problem)

  -- Including identity of "minimum set" of fulfilling projects
  -- Done date : ?

------------------------------
Project list (unordered)

* Control interface in class & port drivers (set/sensemode)
* Enhanced error logging & bugchecks (make consistent, make them
		report what happened)
* Vax?
* White papers re how drivers work & cheatsheets
* Clean up data structures & interfaces
* Enhanced common ring buffer code (trace buffers) to capture
		logging info; in production drivers, controlled via
		control interface
* Big LUNs & IDs
* Bus reset cleanup
* Use of progressive resets (ie, SCSI cmd cancel via abort tag, abort,
	bus device reset, scsi bus reset, in that order and stopping
	as soon as possible)
* Consistent documentation of mount verify interactions
* Documentation (user docs? Design docs?)
* Make queue manager optional for ports that don't need it
* Reorganize class drivers to isolate device specific stuff & facilitate
	response to SCSI capabilities