2-6th October, 2000, York, UK
eXtreme Programming and Modeling: bridging the gap
Marko Boger, University of Hamburg
HGH before and after
'Travel Light' is one of the principles of eXtreme Programming, an
approach to development which provides early and measurably good results
in small to medium sized projects. Proponents of XP eschew documentation,
arguing that it will get out of sync with the code. On the other hand
supporters of a more 'ceremonial' approach assert that specifications,
interfaces, and architecture must all be agreed, documented, and kept up
to date, with coding happening almost as a by-product of these
Given the popularity and arguable virtues of each of these approaches,
how should we choose between them? Are there cases where we can usefully
integrate the two? In this workshop, we will examine issues including:
- On what principles are XP and modeling approaches agreed?
- The role of modeling in RAD: is it for programming in the abstract,
or just for analysing requirements?
- Is modeling more useful for certain design roles than others --
for example, defining component interfaces?
- Which is faster and more effective at finding requirements bugs:
building a rapid prototype, or writing a model?
- Should programs or test harnesses be generated from models?
- Is OCL a better way of writing upfront tests than programming language?
- Test definitions: in component based development, we need to write
interface specifications for components that don't exist yet. Should a chief
architect write these specs in UML/OCL, or as executable test harnesses?
- Partial views: in a large project, different stakeholders have different
requirements of the same use-cases --- for example, safety requirements and
security requirements superimposed on a basic function. Modeling techniques
can handle the composition of these views straightforwardly. Can progamming
techniques do the same thing?
- Can the two approaches be synthesised? In this context, we will look at
Extreme Modeling, a software development process created by Marko Boger and
- In the light of discussion of all the above, what tools emerge as the most
The list of issues will be enhanced by the participants on the website before the meeting using a 'Wiki' discussion< site for the purpose.
Call for Participation
If you would like to participate, you should have some aquaintance with
(and preferably, practical experience with) both UML-based methods (such
as RUP) and XP.
Please prepare a position statement outlining your approach to the issues
raised above, and indicating your experience in the area. Position
statements should be no more than 500 words and should be
mailed to the organisers,
Alan Cameron Wills and Marko Boger
You will be asked to read the other accepted position
statements on the website before coming to the workshop,
and preferably to participate in the discussion forum there. For discussions in advance of the workshop we have set up an interactive discussion site (Wiki) that can be accessed at http://www.trireme.com/scripts/wiki/xpm
All attendees at the workshop will have to pay the workshop registration fee. A discounted fee will be offered for those also
attending the UML 2000 conference.
The session will be organised in such a way as to produce documented
results, which will be made publicly available. Writeup will be
delegated to some of the participants: you are invited to bring along
your own computing machinery.
Format, issues, and example problems may be discussed at the website discussion forum in advance of the workshop.
We will spend the day in a number of timeboxed cycles of presentation of issues and discussion by participants.
We will begin by considering what modeling is for, and how it can fit into various software development scenarios. We will also consider what it is not good for!
For the purposes of focussing discussion, the organisers will introduce a small selection of commercial projects; workshop groups will play the role of methodologists and architects for the projects.
In the morning, we will focus on coming to a common understanding of the issues. While many of us (including the organisers) may have different views about the best answers, we should first explore the problems we're trying to address. In the afternoon, we can move towards discussion of specific syntheses.
Important Dates and Deadlines
July 15, 2000
Deadline for Application to be submitted to
Alan Cameron Wills and Marko Boger
July 31, 2000
Notification of acceptance
October 2-3, 2000
Workshops and tutorials
October 4-6, 2000
UML 2000 Main Conference
Marko Boger has been research assistent for over four years at RWTH
Aachen and the University of Hamburg. He is author of the book Java in
Distributed Systems (in german, to appear in engl.) and chief developer
in the open-source UML-tool project ArgoUML. As member of the
organisation committees of the conferences Trends in Electronic Commerce
(TrEC´98) and Communication in Distributed Systems (KiVS 2000) he has
experiences in organizing and leading conference events. He has held numerous
courses and tutorials in highschool and industry. His main areas of interest are
software engineering and programming for distributed systems, e-Commerce
and internet technology.
Marko Boger and Frank Wienberg are
authors of the Extreme Modeling
Alan Cameron Wills is joint author of the UML-based Catalysis approach to
software development. He is a consultant based in Europe, and over the
past decade has assisted clients in a wide range of application areas. He
is a principal consultant with Trireme International Ltd, whose
consultants introduce clients to a variety of new methods of software
Alan has organised workshops and tutorials in many conferences since
Wiki pages< for this workshop
eXtreme Programming< - for more information on XP
eXtreme Modeling - a web site explaining XM
OCL - a reference and link collection maintained by Jos Warmer, inventor of OCL.
UML 2000 Conference in York
XP 2000 Conference in Cagliari
Hamburg University - more about Marko Boger
Trireme Ltd. - more about Alan Cameron Wills