[alf-dev] Studying the ALF project design rationales for the purpose of

[Daniel Gross <gross@xxxxxxxxxxxxxxx>]

Dear list members,

I am a Ph.D. student at the Faculty of Information Studies at the 
University of Toronto. For my thesis research I am studying design 
reasoning during software design.

I am posting this message to inform ALF project participants of my 
interest in studying the ALF design effort and gathering and modeling 
design rationale knowledge for my thesis research.

I am in particular interested in design situations when there are:

1. multiple stakeholders, who are involved in, and/or are affected by, 
design decision making; and where
2. stakeholders may operate out of separate jurisdictions, without the 
existence of one central authority; and where
3. stakeholders relate through a (social) network of (more or less) 
common interests

In such cases, where decision making is distributed, there is usually no 
one correct answer (design solution) that satisfies everyone, but 
stakeholders need to understand the implications of design options, 
choices and tradeoffs they, and others, may make, in order to achieve 
their respective business and system design objectives.

The purpose of my research is to develop a modeling technique that is 
based on the (ontological) concepts of Actors and Goals. Actors 
represent autonomous decision makers during software system design, and 
Goals represent the Actor's particular motivations or interests, such as 
functionality and quality design goals.

A key objective of the framework is to provide support for capturing and 
analyzing the network of dependencies on design contributions of others, 
which Actors establish and evaluate, when working out design issues, and 
while considering how to best achieve their individual design goals.

The ALF project is particularly interesting for my research, since its 
design effort occurs within the sphere of influence of stakeholders 
belonging to different organizations, having diverse goals, and who are 
involved in, and/or affected by, design decision making. Stakeholders 
include ALF users, service flow designers, applications and tools 
vendors, other middle-ware providers, as well as various ALF requirement 
analysts and architects taking on design responsibility for different 
parts of ALF.

ALF is also interesting since it is an open source project where common 
interests of voluntary participants usually guide the design process. 
The open nature of the project also makes it easier to collect research 
data, since usually all information made available is in the public 
domain. As the ALF project is governed by the Eclipse Website Terms of 
Use (www.eclipse.org/legal/termsofuse.html) none of the ALF material 
available online is private or confidential.

Please let me know if you feel otherwise about particular information or 
documents.

During my study I want to collect and analyze the information and 
documentation found on the ALF website, postings archived in the mailing 
lists and newsgroup, as well as study discussions in conference calls. I 
also plan to ask for consent from the conference call sponsors as well 
as regular conference call members, to record these calls, so that I 
won’t miss any important design discussion during my analysis. I would 
also participate in conference call and may from time to time ask 
clarification questions, but may also contribute to the discussions.

To give you a sense of what I would be doing with the collected data:

 a) I would identify and select particular design issues or concerns 
that are discussed, that exemplify the kind of distributed design 
reasoning I am interested in.

 b) I would collect relevant design reasoning discussions from the 
collected data and relevant external references.

 c) I would then analyze this material to develop a modeling technique 
to capture particular lines of reasoning and argumentations that are 
discussed.

 d) I would also want to post sample design rationale models onto the 
mailing list and/or newsgroup, to share with you the kind of models I am 
developing.

To establish my modeling technique I would for example look at why, and 
in what way, ALF designers intend to depend on the design solutions, or 
the design solving capabilities, of others, in order to achieve their 
design objectives, and in what way such dependencies would be viable, 
given the design objectives of others.

For instance, I could be looking into how ALF designers considered 
dealing with security and authentication issues when integrating tools 
of multiple vendors. Security related design discussions are not only 
about technical possible solution, but also about how feasible or viable 
proposed solutions are for different types of vendors in light of their 
respective motivations and needs.

To study the modeling approach in the context of a project like ALF 
would be very important for my research, as it would allow me to develop 
and test my proposed modeling techniques in a real-world design setting, 
and enable me to get feedback from project participants on the utility 
of the modeling and analysis techniques.

The study could also benefit the ALF project, as I would produce models 
(diagrams) that would summarize and help elucidate, given particular 
lines of design reasoning, how and how well stakeholders are able to 
achieve their design objectives. During the case study contributors to 
the ALF project could make use of these models during or after their 
discussions, or even consider enhancing the models produced and/or do 
their own modeling and analysis of design argumentations. Any such 
contributions would be most welcome.

Best regards,

Daniel Gross
Ph.D. Candidate
Faculty of Information Studies
University of Toronto
Email: gross@xxxxxxxxxxxxxxx

Eric Yu (Supervisor)
Associate Professor
Faculty of Information Studies
University of Toronto
Email: yu@xxxxxxxxxxxxxxx