A personal view of SigAda

[rod.chapman@praxis-cs.co.uk (Rod Chapman)]

SIGAda 2002 - A brief report
----------------------------

Here are my thoughts on SIGAda 2002, held in Houston last
week.  Please note these are purely my personal thoughts
and impressions.  They only reflect a subset of the
papers that caught my particular attention.  There were a
couple of sessions and many tutorials that I missed altogether.
I leave it to others to fill in the blanks.

Overall
-------

Generally, a good SIGAda.  Much credit to all the SIGAda
organising committee and program committee for such
a well-run event.

Turn-out was OK - about the same as the previous years
at JHU and Minneapolis I think.  It addition to all the
"usual suspects", there were a few notable new faces such
as NASA, TRW, Northropp Grumman, and Lawrence Livermore
Labs.

Tutorials
---------

I attended Michael McEvilley's tutorial on the Common
Criteria.  Good stuff - the CC is an important standard
that the Ada community needs to be aware of.  The ComSec
world also appears to be a market where Ada could make a useful
contribution.

I also gave a full-day SPARK tutorial.  Very tiring!

Keynotes
--------

A particularly strong set of key-note speakers this year.

The keynote from Robert Carey of Livermore Labs was amazing - they
are building a thing called the "National Ignition Facility" (NIF),
which is basically one of the world's biggest lasers.  This is one of the
last great "big physics" experiments left in the world.  It will
be used to generate small nuclear fusion experiments and so on.

Check out www.llnl.gov/nif

Some metrics:

 Peak power output: 500 Terawatts
 Pulse energy: 1.8 Megajoules
 Shot time: 3 - 20 ns
 Laser Amplicification gain: 3 million billion (!!!)
 
The control system is a large distributed network of 
machines - mostly off-the-shelf PCs, networking kit and
embedded processors with some custom-built hardware.  They expect
it will be about a million lines of Ada and Java when finished.
Most of the GUIs are Java, with the main control system being Ada.

Oh, and when the shot fires, the whole thing is supposed
to be synchronised to <20 _pico_seconds!  Now that's what I call
hard real-time!


Robert Dewar spoke about Open Source, Free Software, development processes
(i.e. cathedral vs bazaar) and its effect on software reliability
and security.  Robert made the important point that "Free" or
"Open Source" licensing and "Open" development processes have
absolutely nothing to do with one another!  Lively and topical as usual.


The keynote from Charles McKay was very good.  Basically, he
spoke passionately and strongly in favour of remembering
Ada's "roots" (i.e. mission and safety-critical systems) in
the run up to the Ada0Y revision process.


Michael McEvilley did a good talk about the Common
Criteria, security stuff, and why the Ada community has something
to offer in the ComSec world (i.e. we know how to build really
reliable software!)


Papers
------

Some good stuff.  Some stuff I missed.  Usual stuff from
the vendors.  Two highlights for me were:

Richard Conn talked about the development processes used on
the Lockheed Martin C130J software systems.  Their development
process is now very mature (they have a CMM Level 4) combining
all sorts of best practices.  The metrics presented by Richard
suggest that the software development process is now one of
least error-prone activities in the aircraft's development.

VDot Santhanam from Boeing Wichita presented their "ZBra"
Ada subset and compiler.  This is an Ada subset, compiler
and virtual-machine that is designed to be qualifiable as
a development tool to the standards required by DO-178B Level A.
This is a big breakthrough - a level A qualified compiler
has never been attempted before to the best of my knowledge, so
this is a significant announcement.


Next Year
---------

San Diego!  Excellent - let's hope the weather is better - well..
couldn't be much worse... :-)

That's pretty much it...like I said, I hope someone else
can fill in for the bits I missed...

 - Rod