 |
|
|
|
Name: |
Dr. Egbert-Jan Sol |
| Company:
|
TNO Science & Industry, The Netherlands |
| Function:
|
Chief Technology Officer
|
|
|
Egbert-Jan Sol (1956) is CTO (Chief Technology Officer) at TNO Science and
Industry (TNO Industrie en Techiek), located in Delft and Eindhoven. TNO is a 4500
people large contract Research organization in the Netherlands, TNO Science and
Industry is the 1000 technology division of TNO.
Before joining TNO he was vice president Technology of Ericsson-Netherlands and
spent 4 years on assignment in Sweden at LM Ericsson, Corporate Technology.
Egbert-Jan Sol has an industrial career starting as system architect, project manager
and software development manager. Latter he became consultant, marketing
manager and manager of a 500 person R&D centre.
From 1990 till 1998 he was part time professor at Technology
Management at the Eindhoven University of Technology.
He was co-founder and first chairman of the Digital City
Eindhoven (www.dse.nl).
He is also a regular speaker and columnist on new technologies,
in particular on the longer term evolution of the technology.
“If you understand the third generation Internet “the Mobile Internet, with a billion
users beyond 2005, then the 4th generation Internet “the embedded Internet” with the
100B ambient intelligent device around us becomes visible around 2010”
“It is very easy to make technology complex, it is very difficult to make technology
simple. Yet only the simplest workable solution becomes a success.
“The fifth Kondratieff wave, starting around 1990 with cheap computers, booming by
2000 with cheap communication, is not ending yet, but will continues till the 2010-
2025 with cheap and usable devices into the Internet of Things with all kind of
intelligent cubic inches devices around 2010 en push-pin sized devices by 2025”.
|
|
|
|
The technology development is easily predictable. However, the speed and the measure of success are more difficult to predict.
It is interesting to understand which technology develops and in what direction. But, the point is whether you can create
value out of this. Begin early, figure out what will be the critical factors of success, understand the game rules, and
move yourself into a strategic position and develop the right client relationships. As you can notice, these subjects are
not technical, but are more or less economical, marketing or management related. The game between understanding
the technology and applying it is the red-line of this presentation. The focus is put on the ever-smaller products, the so
called intelligent ‘punaise’ and ‘pleisters’, that are so small such that they cannot me assembled by human labor.
Therefore, the challenge lies in taking a strategic position in the development of machines to produce such products.
Actually, this is also the essence of the strategy of TNO Industry and Technology. We try to make choices to invest in
technological knowledge within the kinds of industries that have a chance to survive the global competition in the
coming years and that have a change to become European and worldwide knowledge intensive industries. Examples
are the Holst Center of TNO, IMCE at the HTC, and the concentration of automotive research in Helmond with the
attention on the automotive control systems.
Back to the Program Page
|
|
Name: |
Scott W. Ambler, MIS |
| Company:
|
IBM Rational, Canada |
| Function:
|
Practice Leader
|
|
|
Scott W. Ambler is the Practice
Leader: Agile Development in IBM Rational’s Methods
Group and is the founder of the Agile Modeling (AM), Agile
Data (AD), Agile Unified
Process (AUP), and Enterprise Unified Process (EUP) methodologies.
Scott is the
(co-)author of several books, including Refactoring Databases,
Agile Modeling, Agile
Database Techniques, The Object Primer 3rd Edition, and
The Enterprise Unified
Process. Scott is a columnist with Dr. Dobb’s Journal.
His personal home page is
www.ambysoft.com/scottAmbler.html. |
|
|
|
Many people will tell you that you can't be agile if you're working with a large team, if you're working on a dispersed
team, if you're working with legacy data sources, or if you're working in a regulated environment. Those people are
mistaken. The principles and practices of Agile Modeling (AM) can be tailored into your agile process to specifically
address the needs of "less-than-ideal" situations. In this presentation we'll discuss why up-front modeling is a
risky strategy; how to instead model in a streamlined just-in-time (JIT) manner which captures the benefits of modeling;
how Agile Model Driven Development (AMDD) and Test-Driven Development (TDD) provide significant synergies to
your team; and why creating "just good enough" models, designs, and documents is the most effective strategy
available to you in any area, including the area of embedded systems.
Back to the Program Page
|
|
Name: |
Dr. Pierre van de Laar |
| Company:
|
Embedded Systems Institute, The Netherlands |
| Function:
|
Research Fellow
|
|
|
Pierre van
de Laar recently joined the Embedded Systems Institute
(ESI) as a Research Fellow. He worked 8 years at Philips
Research on exploiting architectural description languages
in component-based software stacks for product families.
He exploited architectural description languages to visualize
and verify the software stacks. Furthermore, he used them
as a basis for code generation and the addition of aspect
orientation to the existing software stacks. Pierre van
de Laar received his PhD from the Catholic University
of Nijmegen. |
|
|
|
Companies develop large numbers of embedded systems that contain a significant amount of software. The amount of
software in these embedded systems is exponentially growing according to Moore's law. With the increase of software
also the complexity increases. Separation of concerns, i.e., the ability to deal with the difficulties, the obligations, the
desires, and the constraints one by one [Dijkstra, 1976], is needed to cope with this growing complexity and to ensure
that companies can continue to differentiate with software.
Currently, embedded software is modularized based on
functionality. Unfortunately, this kind of modularization
cannot separate all concerns as can be observed in the
current software:
» Many (non-functional) concerns are not localised in one software module but are scattered throughout the software.
» Multiple concerns are tangled in one software module.
Throughout the whole software development process, the impact of a limited separation of concerns is noticeable and
leads to problems. To solve these problems, we ask ourselves the question: how can we improve the effectiveness of
our modularization?
Aspect orientation is a technology that improves the effectiveness of modularization . During the presentation, I will
both explain aspect orientation and show using some example the advantages of aspect orientation is practice.
Back to the Program Page
|
|
Name: |
Prof. Dr. Uwe Assmann |
| Company:
|
TU Dresden, Germany |
| Function:
|
Professor
|
|
|
Uwe Assmann holds the Chair for Software Engineering at Technische UniversiteitDresden, Germany. He received his PhD
from Universiteit Karlsruhe in 1995, showing how to generate program optimizers from graph rewrite systems (transformation tool OPTIMIX). During the Esprit project COMPARE he contributed substantially to the design of the CoSy compiler component system, marketed by ACE bV, Amsterdam. From 1995 to 1996 he worked in a post-doc year at INRIA Rocquencourt in the Attribute Grammar group. In the last years, he has developed an invasive software composition system, COMPOST, for grey-box component composition. The system is currently being generalized to components of specification and modeling languages, including UML-XMI. In 2003, he published the book "Invasive Software Composition", which develops a unified technology for generic, connector-, view-, and aspect-based programming. In 2001, he became lecturer at Linkopings Universitet, Sweden, and full professor in 2003. Since mid of 2004, Uwe Assmann holds the chair of Software Engineering at Technische Universiteit Dresden.
One focus of the group in Dresden is model-driven development (MDD). In June2004, Uwe Assmann organized one of the first European workshops on "MDA, Foundations and Applications" (MDAFA 2004). In 2005, this workshop has joined forces with others to the "European Conference on MDA - Foundations and Applications" (ECMDA-FA.org). He has served as project manager for several EU projects (COMPARE, EASYCOMP, JOSES, HIDOORS, REWERSE). In HIDOORS, MDA has been used as a foundational technology for model-driven development of real-time systems. In REWERSE, component models for ontology languages are developed. Uwe Assmann's tool OPTIMIX has also been applied to model transformations in the context of UML. Uwe Assmann is a member of the IFIP working group 2.4 (System Implementation Languages) and the German Computer Society (GI).
|
|
|
Model-driven development and component-based software development are approaches to product-lines, in which software artifacts, both models or code are reused thoroughly. However, the manner in which skeletons of applications (here called PIMs, platform-independent models, or DSMs, domain-specific models; there called frameworks) are instantiated towards applications, differs enormously. While PIMs are translated towards applications, components are linked, composed, or connected.
Is there a way to combine both approaches? How to embed components into MDD, i.e., how to build, design and use MDD components?
In the last years, our group has found a way to build fragment-based component models for every language. Given a metamodel of a language L, a component model can be systematically generated for L, so that a reuse-language results, in which fragment components can be composed. Since this principle is universal, component models for modeling and specification languages come for free and the way to a UML component model is no longer far. With such a component model, many interesting UML-component-based compositions come for free: semantic templates, semantic macros, views, mixing layers, and aspects. Since the underlying tools are universal, this paves the way for true MDD components.
Back to the Program Page
|
|
Name: |
Dr. Sagar J. Chaki |
| Company:
|
Software Engineering Institute, USA |
| Function:
|
Senior Member of the Technical Staff
|
|
|
Sagar J. Chaki is a Senior Member of the Technical Staff at the Software Engineering Institute (SEI), USA. He is a member of the Predictable Assembly for Certifiable Components (PACC) team. His responsibility includes assisting in the development of the Component Formal Reasoning Technology (ComFoRT) as part of the PACC initiative. ComFoRT is a reasoning framework for predicting whether a system will satisfy its safety, reliability, and security requirements. In ComFoRT, these requirements are encoded as behavioral assertions that are verified automatically using exhaustive model checking technology.
Sagar received his Ph.D from Carnegie Mellon University,
Pittsburgh, USA. His research interests include formal
methods, specification, verification, analysis and testing
of software systems, concurrency and software security,
and publish-subscribe systems. Further details of his
research projects and publications can be found on his
SEI staff page at www.sei.cmu.edu/staff/chaki.
|
|
|
Model checking is an automated technique for verifying that a system adheres to temporal logic specifications. Over the last few years, model checking has been applied successfully for finding flaws in critical software systems such as Windows device drivers, Linux, and Firefox. However, the use of model checking for certifying the correctness of systems has been relatively limited for two important reasons. First, a model checker does not emit a proof when it claims the correctness of its target system. This forces us to treat the model checker itself as a Trusted Computing Base (TCB). However, state-of-the-art model checkers are themselves complex software artifacts and should not be trusted any more than the systems they analyze. Second, present-day software model checking technology operates on source code and is unable to target low-level programs. Once again, this means that we have to trust the compilers that translate source code to machine code.
In this talk I will present an approach that removes both the software model checker and the compiler from the TCB. In the first step, we use a certifying software model checker that accepts a high-level program and a specification and, if the specification is found to be valid, generates certified source code. In the second step, we use an off-the-shelf compiler (such as gcc) to transform the certified source code to certified machine code. The certificates emitted as part of the end product can be validated without having to trust either the model checker or the compiler. I will present a practical demonstration of this technology that accepts an UML Statechart annotated with temporal logic specifications and generates certified PowerPC assembly code.
This work represents an ongoing Independent Research and Development Project at the Software Engineering Institute (SEI) and involves members of the SEI's Predictable Assembly from Certified Components (PACC)
initiative -- Kurt Wallnau, James Ivers and Sagar Chaki -- as well Prof. Peter Lee and his group at Carnegie Mellon University.
Back to the Program Page
|
|
 |
|
|
|
