🚀 go-pugleaf

------- Forwarded Message

You are invited to attend the April Meeting of DC SIGAda. After a two
month hiatus, we will return to our regular meeting site, the McLean Hilton.
Details are as follows

Date:	Wednesday, April 15
Time:	7:00pm Refreshments
	7:30pm Meeting
Place:	McLean Hilton
	7920 Jones Branch Road
Sponsor:	Verdix

Topic:                An Asyncronous Life-Cycle for Ada
Speaker:	Gil Prine, Software Manager, MCCI

At last month's meeting, the panel identified Object Oriented
Development (OOD) strengths and application issues.  April's
discussion will be a real-life approach that addresses many of
the issues by integrating four asynchronous processes:

o    domain analysis with "active boundary model" prototyping
o    recursive OOD "analyze a little/design a little/code a little"
o    continuous "iceberg" implementation and capture of objects
o    formally tailored DOD-STD-2167A documentation and controls

into a sum that is greater than its parts.  The need for formal
requirements specification and design definition (DOD-STD-2167A or
equivalent) is acknowledged, but the primary focus is on effective
customer interaction, early prototyping and continuous
implementation.  A single standard is established to control the
development and capture of objects in a reuse pool for all phases
of development.  A program data base provides the formal control
and enhanced visibility necessary to use the recursive "analyze a
little/design a little/code a little" paradigm of Ada OOD.  This sums
up to an effective alternative to the strict waterfall requirements
cost accounting philosophy that is pervasive in software engineering
and is perceived by many as being mandated by DOD-STD-2167A.

The model for the waterfall life-cycle comes from hardware
engineering.  Traditionally, the waterfall splits development into
disjoint efforts at the end of the requirement specification and
design.  This does not work any better for hardware than for
software, and the DARPA Initiative in Concurrent Engineering (DICE)
seeks to minimize hardware life-cycle discontinuities by improving
communications with manufacturing and logistics early in the
design.  In addition, there are significant differences between software
and hardware, especially in implementation.  The marginal quality of
"quick and dirty" prototype is not as apparent as wires hanging from
a breadboard and a working prototype and can be perceived as a final
product by hard-pressed program managers and/or eager customers.
Software development should use Fourth Generation Language (4GL)-like
active boundary model tools to create production-quality objects as
early as possible in "quick and dirty" prototypes to exploit software
characteristics.  This approach and its importance as the cornerstone
of the Asynchronous Life-Cycle will be discussed in depth.

Mr. Prine has thirty years experience as a programmer analyst,
systems engineer, and manager.  He is currently responsible for
Ada development as Software Manager at MCCI.  As Technical Director
at Andrulis Research Corp., he developed and used the Asynchronous
Life-Cycle on a 200K LOC Ada Navy project.  Mr. Prine has been
involved in Ada methodology R&D since being an invited speaker at
the first Ada Environment Workshop in November 1979 and gave a DC
ACM 1992 Spring Series of Professional Development Seminars on
the Asynchronous Life-Cycle.  He has an MS from Johns Hopkins
University, and has taught Ada at Montgomery College since 1989.

So mail off those tax forms and come join us at the McLean Hilton.

Dave LaVallee
Program Chair

------- End of Forwarded Message