To:   Distribution
         From: Keith Ball
         file: c2meet2.text
         date: 6-January-1984

         Re:   Second Constellation II meeting

         The second Constellation II meeting was on January 5,
         1984 at 3:00 to 5:00 pm.  Present were : Phil Belanger,
         Sean Finley, Bruce Kendall, Denise Pitsch, Steve
         Schwartz, Robin Yonash, and Keith Ball.

         Denise presented her boot procedure document.  Three
         unresolved problems were specified. One, the lookup
         protocol must have generic as well as specific device
         types for all devices.  Two, what is an adequate timeout
         lengths for waiting for response from disk servers after
         broadcasting a message.  Denise assumed 5 seconds.  The
         Concept uses approx. 14 seconds.  Three, we need a new
         version of the H-drive firmware to be implemented with
Phil     all the new drive commands.  After Phil resolves the
Denise   first two problems Denise will be able to finish the
         document.

         Steve informed us that a bug exists in the MSDOS 2.0
         logon program for the IBM PC which prevents the user
         from logoning onto a UCSD user name.  This bug must be
         fixed when we next work on the PC.

         I discussed the physical transporter driver and its
         relationship to the rest of our system software.  For
         this development project we will only build the physical
         transporter driver, with buffer management, and not the
         socket management portion of NIS.  The physical
         transporter driver will hide the existence of the
Keith    buffer.  My main action item is to complete the physical
         transporter driver specification.  Robin was concerned
         on the impact of this change to the SNA gateway and
         other software developers.  We must have information
         available to them as early as possible.

         We agreed that when the physical transporter driver is
         loaded into the PC then the transporter rom will not be
         used.  All communications with the transporter will be
         through the physical transporter driver.  However, the
         rom will be used for booting.

         Bob W. gave me a new proposal for volume locking similar
         to the method Denise has used on the Product Development
         net to lock shared volumes.  We agreed this was the
         simplest and best method of the three documented.  I
Keith    will write-up the document and send it around.

         Denise implemented a version of the spool driver on the
         Concept.  We determined that the default server to spool
         to is the "home server".  Steve was concerned about the
         OmniDrive losing the pipes on reset.  Phil said he would
Denise   check into it.  Denise volunteered to specify the
         program to close the pipe and set the driver options.

         For the interrupt IBM transporter, Steve was concerned
         that the full 4K of eprom on the card is not fully
         decoded.  I will check into it with Bruce E.  In order
         to release a new version of the tranporter rom to fix
Bruce    bugs and add some minor enhancements, Bruce K. will
Steve    write-up the changes needed and Steve will implement
         them.  Steve said most of the changes have been
         completed for the Companion.

         Additional action items are :

         Denise will specify and schedule, as much as possible,
         the changes necessary to the system and user utilities
         and the boot process;

         Bruce K. will determine if the work can be performed on
         the SONY system;

         Bruce and I will specify and schedule the driver work,
         disk, physical transporter, and spooler.

         Distribution

         Phil Belanger
         Sean Finley
         Bruce Kendall
         Denise Pitsch
         Steve Schwartz
         Robin Yonash
         Bob Wohnoutka