Dagstuhl Seminar #03061 - Sofware Architecture Recovery and Modelling

Dagstuhl seminar 03061

This page describes the submission guidelines for the Dagstuhl seminar 03061 Software Architecture: Recovery and Modelling organized by Arie van Deursen (CWI Amsterdam, The Netherlands), Rick Kazman (CMU Pittsburgh, USA), and Rainer Koschke (Uni. Stuttgart, Germany).

Collaborative architecture reconstruction and modeling task
Preliminary programme
Submission of abstracts for presentations
Submission of abstracts for Special Issue in Journal of Automated Software Engineering

Collaborative architecture reconstruction and modeling task

SWARM Architecture Recovery Exercise - Duelling Web Servers

We propose doing a joint/parallel architecture reconstruction activity at the Dagstuhl SWARM workshop. The idea of the activity is this:

We have three systems that we will work on. We will bring our tools and expertise to bear on reconstructing these systems. The systems chosen for this exercise are all Web servers: Apache, Roxen, and Jigsaw.
We will divide ourselves into three sub-groups, and each sub-group will work on one server (either collectively or individually)
We have a published reference architecture for Web servers, which we can use as a jumping-off point (and initial architectural hypothesis) for the reconstruction tasks.

Given this basis, we would like everyone to try to reconstruct their system. Different people will have different tools or methods that they can bring to bear on this problem. For example, one tool might be able to extract the based set of "facts" that describe the system at the code level. Another system might extract build information about the system. Another system might be able to determine run-time information about the system.

Things to do before you come to Dagstuhl

Before you come to Dagstuhl, please download and read the Hassan/Holt paper on the reference architecture for web servers. Moreover, please download the source code of the following web servers and apply your tools to the systems (at least, to one of them). You are not asked to do a deep analysis, however; just make sure that you will be able to analyze them at Dagstuhl. It would be great if you could share any intermediate result with other attendees. To that end, we created a Wiki that allows you to share information with other participants ahead of the Dagstuhl seminar (things like extracted facts, visualizations, or design documents).

Things to do at Dagstuhl

Sub-group members may work individually or together to reconstruct the most complete, most correct, most justifiable version of their server architecture that they can. Different teams may have different (even competing) working hypotheses about the architecture of their server and may choose to collect evidence to support their position. Teams may also choose to look for violations in the reference architecture, and to understand the reasons (if any) for these violations. This will help: 1) update the reference architecture; or 2) feed the analysis of each individual server architecture (to understand how it aligns or does not align with the reference architecture, and why).

We look forward to this exercise having both collaborative and competitive components. We encourage teams to collaborate with each other and learn from each other. We hope that teams will share their knowledge and their fact base so that the collective experience is greater than the experience of any individual team.

Preliminary Programme (Version 3)

MONDAY 3 FEBRUARY

08:45 Opening and Welcome 09:00 Session 1.1 Specification and Modeling C. Hofmeister and R. Nord. Software Architecture Views for Enterprise Applications Amnon H. Eden. Two-Tier Programming Liam O'Brien Aligning Architecture Reconstruction with Existing Architecture Documentation Approaches 10:30 Coffee Break 11:00 Session 1.2 Components Mike Godfrey Modelling and Extracting the Build-Time Architectural View Thomas Eisenbarth and Daniel Simon Recovering Static Component and Connector Views on Architecture Spiros Mancoridis Characterizing the `Vulnerability Likelihood' of Software Components 12:15 Lunch 14:00 Work in teams on virtual "Software Architecture Reconstruction and Modeling" book (see separate description) 15:30 Tea Break 16:00 Virtual Book continued 18:00 Dinner 20:00 Book Festival TUESDAY 4 FEBRUARY 09:00 Session 2.1 The Module View Point Ric Holt Central Role of the Landscape View of Software Architecture Brian Mitchell Visualizing the Search Landscape of Metaheuristic Software Clustering Algorithms Jens Krinke Apect Mining 10:30 Coffee Break 11:00 Session 2.2 The Dynamic View Point Alexander Ran Recovering runtime architecture from execution traces Katharina Mehner Dynamic analysis and modeling of concurrency aspects of components (Immediately after session 2.2) Announcement Scott Tilley and Shihong Huang Introduction to an Experiment to Assess the Efficacy of Software Architecture Visualization Techniques for Recovered Artifacts 12:15 Lunch 14:00 Team Work on Joint Reconstruction Exercise 15:30 Tea & Coffee 16:00 Team Work Continued 18:00 Dinner 20:00 Drinks and Games WEDNESDAY 5 FEBRUARY 09:00 Session 3.1 Consistency, Traceability, and Design Decisions Nenad Medvidovic Coupling Architectural Discovery and Recovery to Stem Architectural Erosion Harry Sneed Reestablishing Links between Semantic Levels in Software Systems J. Bosch Explicit Modelling of Architecture Design Decisions 10:30 Coffee 11:00 Session 3.2 Meta-Models and Data Jean-Marie Favre Architecture Reconstruction using Meta-Models Andreas Winter Towards a Catalog of Reference Schemas in Reverse Engineering Jörg Wadsack Architectural Issues in Data Reengineering 12:15 Group Photo 12:30 Lunch 14:00 Social Programme: Hike 18:00 Dinner 20:00 Tools Demonstrations Bring your own laptop and show your latest tools to new users THURSDAY 6 FEBRUARY 08:45 Session 4.1 User Involvement André Marburger Requirements of Architecture Recovery in the Telecommunication Systems Domain - an Example Hausi Müller Adoption-Centric Software Engineering Jean-Francois Girard Evaluating the Effect of Expert Involvement in Architecture Reconstruction Scott Tilley and Shihong Huang Assessing the Efficacy of Software Architecture Visualization Techniques for Recovered Artifacts 10:30 Coffee 11:00 Session 4.2 Case Studies Andrey A.Terekhov Recovery and Improvement of Software Architecture: A Case Study Arie van Deursen Software Architecture Reconstruction in an Auditing Context 12:15 Lunch 14:00 Teams finish their reconstruction case studies 15:30 Tea and Coffee 16:00 Reconstruction Presentations Team by Team 18:00 Dinner 20:00 Drinks and Music FRIDAY 7 FEBRUARY 09:00 Session 5.1 Software Product Lines Chris Verhoef Software Product Line Migration and Deployment Liam O'Brien Techniques for Mining Software Architectures and Components for Product Line Claudio Riva Architecture Reconstruction at Nokia 10:30 Coffee 11:00 Session 5.2 Tools Kamran Sartipi Alborz: a framework for architectural recovery Jörg Wadsack and Jens H. Jahnke A History Concept for Recovery and Design Tools 12:15 Lunch 14:00 The End

The SWARM Book

In this assignment, you are asked to compile a proposal and table of contents for a virtual book addressing the topic of "Software Architecture Reconstruction and Modeling". You will work in groups of approximately 5 people.

The purpose of this assignment is

To get an overview of the key questions and results of the SWARM research area;
To compare the different taxonomies from all of the groups;
To position your own expertise and interests in the overall field;
To give you an opportunity to learn about your team member's expertise in detail;
To give your team members the opportunity to learn about your expertise;

You are requested to compile:

a global table of contents;
detailed layouts for selected chapters;
a 5 minute (max.) presentation of the book, convincing potential buyers (students, practioners, instructors, researchers) of the benefits, and convincing potential publishers of the profitibality of the project.

All 5-minute presentations will be held in a 2-hour book festival. The various proposals will be juried, and the winning proposal receives ... a secret prize.

Quality Attributes
Software Product Lines
Reconstructing Variability
Remodularization
Visualization
Viewpoint Reconstruction
Components and Connectors
Cross-Cutting Concerns
Reconstruction Processes
Agile Architectures
Software Process
Customer Involvement
Requirements Recovery
Architecture Documentation
Architecture Modeling
Concurrency
Architectural Migrations
Security
Formal Methods
End User Adoption
Patterns
Architecture Evaluation
Software Evolution

This list is not exclusive and fairly arbitrary: you are invited to come up with your own structured set of topics. Make sure you include topics reflecting your expertise and interests!

Naturally, you are free to use your creativity and originality where ever appropriate -- feel free to deviate from specific instructions when necessary to achieve a better result.

Observe that this book project is entirely virtual: there are no specific plans / publisher contacts to actually come up with a book like this. This assignment is merely intended as a tool to help us collaborate and rethink our field.

The proposals will be judged by a jury according to a format also used in the famous European Song Festival. Each group selects one of its members as jury representative. Each jury member ranks the various and gives points to the five proposals he or she likes est: 5 points to the top proposal, 4 to the second best proposal, etc. The proposal receiving overall most points is the winner. A jury member cannot give points to his or her own proposal.

During the presentations, the audience can attempt to influence the jury. For example, you can explain why you think the setup of a certain book is fundamentally wrong, or why the treatment of a particular topic in another book proposal should be considered a breakthrough in bringing architectural principles to the masses. Teams are permitted to adopt a systematic lobbying strategy, but such efforts can be stopped by the session chair.

Submission of Intended Presentation

If you wish to present your work at the Dagstuhl seminar, you are requested to submit a title and abstract before 1st December 2002.

The organizers will use these to compile a program which will be made available in advance. If we have more presentation proposals than we could reasonably integrate in the seminar, a selection of presentations will be made.

Authors are invited to submit an electronic version of their contribution in PostScript, PDF, RTF, or Microsoft Word format to swarm@droste.informatik.uni-stuttgart.de using "SWARM talk" in the subject line. The abstract should be 1-3 paragraphs long, focusing on the key issues addressed in your presentation: What will we learn from your talk (i.e., why should we listen to you)? Based on the abstracts, we will group presentations into common themes. Please provide us with enough details for this grouping. Please indicate the amount of time required for your presentation. Your talk should be in one of the following categories:

5-min position statement
10-min presentation
15-20 min presentation

If we cannot give everyone the time she/he requires, we might need to shorten the time. We will use the abstracts to decide which talks should be shortened. The more detailed your abstract and the more interesting the content, the higher are your chances to get your full time. Moreover, we will put your abstract on the seminar web page. We appreciate your understanding and collaboration.

Submission for Special Issue of Automated Software Engineering

A special issue of the Journal Automated Software Engineering devoted to this seminar will appear. If you are interested in submitting a paper to this special issue, send an extended abstract or full paper before 13 December 2002

Based on the presentations and the submitted extended abstracts we will invite selected authors to to revise, extend, and submit their papers for consideration for publication in this special issue.

The final manuscript is due after the seminar: We would like to give you the chance to work in feedback from the Dagstuhl seminar. We are defining the exact date at the seminar.

Authors are invited to submit an electronic version of their contribution in PostScript, PDF, RTF, or Microsoft Word format to swarm@droste.informatik.uni-stuttgart.de using "SWARM ASE" in the subject line. Manuscripts must be in English, single-spaced, 10 point font size, and 15 pages, respectively, 8,000 words maximum. In addition to the electronic version of the paper, the email submission should include a text-only version of the title, author(s) name and affiliation, abstract, and the name and address (both postal and electronic) for the contact author.

Papers submitted for consideration by the special issue must represent original unpublished work. No version of the paper may be submitted concurrently to any other journal or conference.

Last modified: Fri Jan 31 15:24:16 MET 2003
admin@droste.informatik.uni-stuttgart.de