MEMORANDUM

TO:     Rick Greenblatt
FROM:   Keith Corbett
DATE:   28-Jun-88
SUBJ:   Status report
COPY:   Jim O'Dell, Guy Chevalier

The group's activity over the past few weeks has focused on two areas:
1) the forthcoming software Release 4, and 2) ongoing K-machine
internals documentation.  This is in keeping with our priorities:

        1) Release
        2) Maintenance
        3) Integration and Quality
        4) Documentatation and support

1. Priorities

Peter DeWolf and I have been working together on release 4.  David
Saslav and I, working with Bob Kerns and Jim O'Dell, have concentrated
on the K-Machine/Macintosh interface document.

2. Design / Documentation

A report which lists the current status of each document is attached.

The K/Mac interface document is both a functional specification and an
outline of design issues.  The meetings held this month, both in
Cambridge and Montreal, have been very productive. Although many
complex issues need to be further ironed out, it has become clear
that:

  1) The Cambridge group has made substantial progress on specifying
requirements and design parameters, and

  2) Our group will have a few weeks of work to do on integrating the
code into the LISPM system software.  We cannot proceed on integration
at this time because:

        a) Montreal has to catch up on design clarifications
           made recently, and 

        b) We need further technical discussions, demonstrations
           of work to date, and implementation documentation from
           Montreal.

Our major ongoing need is for more direct interaction between the two
groups.  I have three suggestions for improving communication and
productivity.  All of these methods may, in the end, be required:

   1) A real-time, on-line TCP file transfer mechanism between the two
sites.  Serial-IP, my extension to Lambda networking, requires 3-4
more days of my time before modem connections can work.

   2) The Montreal group has expressed an interest in having me come
up for an extended stay, perhaps one week.  This would be appropriate
just after they resume work (when a bus coupler is available).  I
could work with them on LISP system problems as they arise, and become
familiar with the application for later integration efforts.

   3) After they have a prototype that is ready for testing, the
Montreal group should send somebody down here for one week.  This will
help us with the final integration and reduce the need for prior
documentation (David can help them develop it here).

In addition, the language barrier remains a problem.

3. Software Validation

I have ported the ISI CommonLISP code validation software and the
CommonLISP test suite.  The test results provide some measure of our
conformance to "vanilla" CommonLISP.  I have also begun a ZetaLISP
test file for verifying standard features.  The suite is an ongoing,
expanding concern: the tests must be run periodically, for example,
after we have made a new cold load.  This process will alert us to
certain kinds of problems that may be introduced with system updates. 
                                    
4. Release 4

The group has made numerous fixes to various system components,
including low-level utilities such as the compiler, ZMail, ZMacs,
CommonLISP, the network, and the file-system.  Many of the fixes were
motivated and facilitated by a new in-house tool, the validation /
testing system.  This required substantial effort these past few weeks
to fix and improve (see below).

The last "Top Priority" bug fix was made yesterday, and the code has
been "frozen".  We will begin recompiling the system today, and then
construct a new cold load.  We will double-check the software simply
by using it for 3-5 days and re-running the test suite.  If all goes
well, we will be able to make master tapes early next week.

I have discussed the final phase of release activities with Clovis
Regis and Guy Chevalier.  Both Manufacturing and Administration will
be needed to handle production and distribution.

5. Internal Support

Peter and I both spend time almost daily assisting other staff members
with various problems on the Lambda.  This is a worthwhile aspect of
our group's role; aside from the primary responsibility for
maintenance and integration, we do our best when called upon to
provide internal assistance.  Happily, with the presence of Steve
Haflich and Bob Kerns, there are more senior programmers to go around.

6. Resources and Problems

My own efficiency has been somewhat hampered recently by the fact that
David and I must share one Lambda 2x2 with ongoing problems.  The slot 0
processor I use crashes intermittently, and has been down completely
for a total of 4 days in the last 4 weeks.  At this moment I am
waiting for a memory board from Lowell.

I would like to bring one of my systems down from Lowell, which would
be a great improvement, as I need two workstations anyway.  I have
discussed the logistics of doing this with Henry and Clovis.



KMC