Report from the Software Technology Conference in SLC, Utah

Report from the Software Technology Conference in SLC, Utah

[Bert Nelson]

The following is a copy of my draft report of my expereince at the
Software Technology Conference.  This conference has an emphasis
towards Ada usage and software engineering within the defense
industry and the military.

Bert Nelson
bnelson@csulx.weber.edu

with the_usual_disclaimers;


\f






DRAFT
Trip Report
Software Technology Conference
Salt Lake City, Utah
April 19 - 21

Bert Nelson



\f
     On April 19, 20, 21 I attended the Software Technology
Conference in Salt Lake City, Utah.  This report is a summary of
the events that I attended and participated in.  The actual details
of the papers presented will be forthcoming in the proceedings of
the conference that will be mailed out in a few weeks.

     The Software Technology Conference was attended by nearly 1800
people.  These people consisted of representatives from all
branches of the armed forces and the defense industry.  The larger
contractors such as GE and Lockheed had a visible presence at
the conference.

     The conference consisted of eight concurrent tracks including
such topics as software engineering and process improvement.  About
130 papers were presented at the conference, which were taken from
a pool of 500 papers that were considered for the conference.  A
new twist on the conference was its sponsorship.  This conference
was jointly sponsored by the Air Force, Army, and Navy.  In years
previous the Air Force was the lone sponsor of the conference.

     The theme of the conference was "Software -- The Force
Multiplier".  The keynote address on the theme was given by Lt.
General Albert J. Edmonds, Acting J-6 the Joint Staff.  The bulk of
his comments was a proclamation of the continued support of Ada as
the primary language for Department of Defense software.  He said
that "Ada is crucial" and that the military is still committed to
the use of the language.

     Prior to General Edmonds address Lt. General Peter Kind from
the U.S. Army had a few remarks.  He thought that software
productivity is the challenge that we need to face up to.  To
illustrate his concept he told the audience about a study done on
software productivity that included many countries.  The formula
used in this study for software productivity was  
Software Productivity = Cost / Function Points.  There were three
areas included in this study:  management information systems,
systems software and military software.  From this study it was
concluded that in management information systems software the
United States was the leader with a ranking of first in the world. 
In systems software Japan was number one followed by the USA, which
was ranked second.  And in military software France was number one
followed by the likes of Germany and Japan.  The United States was
ranked eighth just above Egypt and Brazil, who were ranked ninth
and tenth respectively.
\f
     In addition to the presentation tracks and the speeches there
were over seventy exhibits by defense contractors and software
companies.  The largest number of displays were in the area of
Computer Aided Software Engineering (CASE) tools.  Most of these
software packages are designed for the workstation environment
running under Unix and X windows.  The largest problem I see with
these tools that if you do use them then you have to make
a large investment in money and time to use them.  Another problem
that comes up is that if a particular package is purchased you may
be locked into a software and the hardware platform it requires for
a long period  of time.

     The most useful tool I saw was an Ada sensitive editor.  This
editor will check for syntax errors and automatically format the
data according the industry standard, which can be adjusted to add
or take away from that standard.  In one example I saw how the word
"goto" was entered on a line. After pressing the return key the
software inserted commented lines that warned the user that using
the goto was against the standard and should be avoided.  The user
has two choices at this point.  He can ignore the
warning and the software will compile since the warning was in the
form of an Ada comment or he can tell the software to remove the
goto will a click on his mouse.

     On the whole the conference was very good.  The organization
of the conference and the presentations were done in a first class
manner.
\f


Buzzwords at the Conference

Object Oriented 
Ancient Artifacts
Real Time Embedded Systems
Domain Engineering
Domain Analysis
Reengineering
Software Reuse
Component Reuse
Reuse
Ada
X Windows
POSIX
Rightsizing 

\f

Summary of a few notes from the tutorials, tracks, and speeches.  


>From the tutorial, "A Reengineering Tutorial" by Evan Lock.

One of the goals of reengineering is the quest to understand how
the software works.  You to facilitate understanding by continually
breaking down the software into manageable pieces everyone can
understand.  And there needs to be a facility (preferably an
automated one) that allows you to retrieve this information
graphically.

Another issue is when you translate old existing code in another
language such as Fortran into Ada.  Do you maintain the code to
fulfill new requirements in the old language or do you go straight
to Ada.   Mr. Lock suggest that "you bite the bullet" and translate
immediately to Ada.



"Software Reuse" by James A. Hess, who works for the U.S. Army

The motivation for software reuse is simple -- the Army currently
has about 43 million lines of code in production and the number of
lines is growing.  There needs to be a method to reduce code and
their associate costs through reuse.

The three driving principles of software reuse are

1) Be Domain Specific
   Focus the software to the task at hand

2) Process Driven
   Reuse should be part and parcel of software development; not an 
   afterthought

3) Life Cycle Component
   Reusable objects are not limited only to code, but can include
   designs, requirements, test plans, etc.

\f

Marie Riggs from the Army Reuse Office

A question from the audience was raised as to the quality of
software in the Army Reuse repository.  She pointed out that in the
Army there are four levels of software integrity involved in
classifying software. The lowest level is level 1, which are Ada
packages that are provided "as is" and may have syntax errors in
them.  The highest level is level 4, which includes a fully tested
Ada package, along with test suites, and documentation for reuse in
the form a reuse manual.

The issue of whether we can we have a central Ada (and other life 
cycle components) repository was brought up.  Ms. Riggs answered by
saying  that the technical problems to overcome in running a
repository are simple  compared to the political and economic
obstacles that come up.


Vice Admiral Jerry Tuttle, United States Navy


One of his main topics was the electronic super highway.  He said 
our challenge is to create software to facilitate the newer, faster
technology. He mentioned that in the last 20 years it has gone from
10 to 7 to 3 years  to get an order of magnitude of increased
performance in computer hardware.  

He mentioned the importance of software in military hardware.  For
example, the new Advanced Tactical fighter will have about 80% of
its functions control through the software and it will take an
estimated 10 million lines of code to accomplish.

His next topic was how to improve the code in system.  He said that
most errors in the code are a result of the requirements phase, but
that only 5% of the errors are caught at this stage.  There is a
need to use tools up front to trap errors, which saves costs.  The
further along you are in the life of a software project the harder
and more expensive it is to fix bugs in the code.  

He mentioned that effective specifications are necessary for good
software.  Some of his final words were that "only software will
unlock the potential of this new hardware".
\f


Phillip Gould from Lockheed 

     Mr. Gould mentioned there is a large issue in the F22 project. 
The largest challenge for the F22 project is to get hardware and
software for the life of the project, which is expected to run to
the year 2033, or about 30 years.  He said that it may be
impossible to support a given software/hardware platform for a
number of reasons, but he thinks there might be an answer.  The
answer is to define clearly the functionality that you want the
software/hardware to accomplish.  By doing this you get what
you want without being tied to a particular software or hardware
platform.


Robert E. Johnson Jr.  "Army Ada Technology Insertion Program".

     The goal of the Army Ada Technology Insertion Program is to
develop software that would go over and beyond current project
needs and would be available for reuse by other projects within the
services, not just the Army.

The idea is to share code that has already been developed and
tested.



"Status of Ada 9X", Chris Anderson, Ada 9X Project Manager

Ms. Anderson gave a very brief overview of the status of the Ada 9X
project. She said that there has been unanimous international
approval of pieces of the Ada 9X project, which has been made
easier by the fact that Ada 83 is an ANSI, Military, and ISO
standard.

Timetable for the completion of Ada 9X

Event                    Projected Date

Draft 9X                 September 1993

Final Approved 9X        December 1994

Validated Commercial     First Quarter 1995
9X compilers available

Ms. Anderson also commented on the growing acceptance by the object
oriented community of Ada 9X.  At one of the big objected orient
conferences later this year an entire section of presentations will
be devoted to Ada 9X.