Session: Industrial Presentations #1 |
|
|
Developing Reliable Software Rapidly |
|
D. N. Kleidermacher
(Green Hills Software, USA) |
Although there is significant
evidence that following a structured, comprehensive quality
management process improves reliability of software relative to
the use of unstructured processes, these rigid methodologies
often cause a loss in efficiency, delayed time to market, and
frustration in the daily lives of software developers and
managers. This presentation will introduce key elements of a
high integrity process that can not only be used to satisfy the
requirements of various high assurance quality standards, but
can also improve efficiency of software development. The process
is presented as a set of 20 guidance statements that are easy to
understand and apply to both legacy and new software development
projects.
|
|
|
|
Publisher Framework (PFW) |
|
J. Klein
and D. Sotirovski
(Lockheed Martin, USA and Raytheon, Canada)
|
One of the lasting challenges
in building distributed fault tolerant systems is keeping
application code size and complexity down. This can be done by
capturing the nuances of distributed computing environment and
redundant fault tolerant elements into a common middleware layer,
thus factorizing the code that would otherwise need to be
written again and again by each distributed fault tolerant
software component. When the application code is loaded with
complexities, and Air Traffic Control (ATC) is certainly one
such example, achieving this goal becomes paramount.
Under a project called En Route Automation Modernization (ERAM),
the Federal Aviation Administration (FAA) is developing a
replacement for its aging en route assets and, at the same time,
creating a foundation for the many future enhancements that the
projected increase in air traffic requires. At the core of the
ERAM design is a distributed object oriented (OO) framework
called Publisher FrameWork (PFW), which is ERAM’s answer to the
aforementioned OO challenge. This presentation describes the PFW
properties, the experiences with it accumulated through the
first build of the ERAM program, and its applicability to fault
tolerant computing.
|
|
|
|
Assessment of Lane Recognition Systems |
|
D. Dickmanns,
H. Graef
and M. Anderschitz
(BMW and Eurospace, Germany)
|
In order to reduce accidents
not only passive safety has to be increased, but also active
safety systems are needed. Active safety systems support safe
driving and inform the driver in an early stage before the
situation gets critical. In order to identify emerging critical
situations the environment of the vehicle has to be perceived.
One important feature to be recognized is the road. With
information on the position of the vehicle in the lane i.e. the
driver can be warned in case of leaving the lane. To detect the
road there are several new sensors under development at the
moment. This paper deals with the evaluation process of lane
recognition systems. The evaluation of the sensor performance is
an important part in the development process and should be easy
to handle. The great challenge here is the huge variety of
environmental situations, from different road constructions up
to all weather conditions. It is not possible to test the
development steps by means of typical test drives. But a full
test is needed because improvements on one side often make other
features worse.
|
|
|
|
Using CORBA to Bring New Life to Legacy Ada
Software |
|
J.-C. Mahieux
(Top Graph'X, France) |
In the context of continuously evolving software technologies,
the question of investing or not in existing Ada software is
often raised. This presentation describes how new innovative
techniques using CORBA can extend the life of an existing Ada
system significantly, with low effort.
The presentation will use as an example a combat management
system developed in Ada.
|
Session: Industrial Presentations #2 |
|
|
A Metamodel-based Approach to Reverse
Engineer Ada Source Code into UML |
|
T. Capelle and X. Sautejeau
(Sodius, France) |
Many companies have valuable
software assets based on legacy Ada source code they want to
reuse in new projects. Very often, UML is another key element in
the development of such projects and the Ada components must fit
in the UML framework that describes the system. This requires
some form of reverse engineering of Ada to UML models. However,
mapping Ada to UML is a challenging task, the main difficulty
deriving from the nature of the two languages. Although they
intersect to a large extent, they are a long way from being
bijective. Being aware that a perfect mapping between these two
specific languages is not achievable, we have to assess what
kind of value can still be obtained from the activity of reverse
engineering. We propose an approach using metamodelization that
refactors Ada code into UML models emphasizing the most relevant
aspects of the system. We implemented this solution in the
I-Logix Rhapsody in Ada CASE tool. In this presentation, we will
first detail some of the main differences between Ada and UML,
then we will review different ways to turn Ada code into
valuable modeling assets and finally we will describe the
architecture of our implementation.
|
|
|
|
The Development and Deployment of a
Workflow System partially written in Ada95 |
|
F. Piron
(KonAd GmbH, Germany) |
In the past 3 years KonAd-GmbH
developed the Workflow Engine KWfl which also contains a
document management component. The workflow server mostly
resides inside an Oracle database as a set of PL/SQL-Packages.
An additional serverside library contains functions used for
document management. This library is completely written in
Ada95. The win32 client software "ElSch" provides a
multithreaded graphical User interface to Kwfl with a look and
feel inspired by windows explorer. ElSch also is completely
written in Ada95.
In September 2005 Kwfl and ElSch were installed in a
governmental institution in the city of Zurich, Switzerland.
In the presentation i will describe our reasons for choosing
Ada95 in this project. The software architecture will be shown
and i will give a short development and deployment history. In
the final conclusion i will discuss our experiences in using
Ada95 to built a Database Application with a rich GUI Part. |
|
|
|
Parallel Graphical Processing in Ada |
|
M. Ward,
S. Palin
and N. Audsley
(BAE Systems and U. York, UK)
|
We have been looking at the
implementation of high-integrity, real-time, image manipulation
systmes. In order to get the performance needed to support high
resolutions and frame rates, a parallel hardware implementation
using the SPARK and Ravenscar subsets was used. The system
provides the flexibility needed for different resolutions and
image processing by allowing the number of parallel tasks to be
changed according to the requirements of the system. This allows
it to run at frame rates of 60Hz, with minimal frame delay, with
good prospects of higher performance. The implementation of the
system brought to light some shortcomings in the Ada language,
and propose the addition of a multi-cast or parallel entry call
semantic, allowing simultaneous task release, or data multi-cast.
|