Sunday, February 17, 2008

Colloquium on "Challenging Groupware: Emerging configurations for distributed interaction"

Gabriela had spotted this colloquium titled "Challenging Groupware: Emerging configurations for distributed interaction" and we agreed that it seemed very interesting from our point of view, both because of the topic and because of the people giving presentations. Eventually it was decided that I would have the privilege to attend the event and I consequently traveled to London and the colloquium venue Commonwealth Club on the 12th of February (incidentally the same date as Gustav III becomes the King of Sweden back in 1771...)

The event was organized by MiMeG Research Node of the UK ESRC National Centre for e-Social Science, including Jon Hindmarsh and Dylan Tutt (both of King's College London). Some of the speakers were Christian Heath (King's College London), Tony Tang (University of British Columbia), Paul Luff (King's College London) and Gloria Mark (University of California, Irvine).

The purpose of the colloquium was to highlight and discuss issues "raised as a result of distributed collaborations, that are becoming increasingly important in the context of social and organisational developments, such as the emergence of global firms and scientific collaboratories."
In his introductory talk, Christian Heath argued for the importance of understanding three aspects in particular in groupware research: peoples mutual perception of objects, supporting participation and co-participation, and the integration of all resources (digital as well as tangible) in the collaboration and interaction. He also made the point that - even today - social scientists in this area know little about the organization of the interaction and that this field is dominated by social challenges as well as technical.

In line with this, the focus of the talks encompassed evaluations of concrete technical solutions (such as Kenton O'Haras overview of HP's HALO Collaboration Environment), more generic conceptual investigations of collaborative interaction (such as Tony Tang's talk about the dynamics of sharing a display in a small group) and macro-perspectives on issues highlighted by groupware research (such as Paul Luff's overview of their research on creating media assemblies and Gloria Mark's talk about the trend of new collaboration paradigms in today's world).

All in all, it was a very interesting event to attend. Although most of the content was not directly applicable for what we are doing, there were many overlapping areas and highlighted issues that can inspire our own research direction. Lastly, I believe that Cambell Catering here at UL could have learned a lot from the Commonwealth Club catering...the light lunch snacks were exquisite, nutritious, and very filling... No greasy sausages or deep fried chicken legs as far as the eye could see! :-P

Wednesday, February 13, 2008

Celebrating Prof David Lorge Parnas at UL

Yesterday, UL organised a lecture (and a reception) to celebrate the IEEE Computer Society's 60th Anniversary Award made to Professor David Lorge Parnas.

Hosted by Prof. Vincent Cunnane, UL's VP Research, the event included two talks, one by David Parnas himself and one by Fionn Murtagh, the new Director of ICT at the Science Foundation Ireland.

The title of Prof. Parnas' talk was:
‘Getting Lucky’ in Software Engineering Research. He spoke about his most important accomplishments and offered a possible recipe for success:

1. Start with a real problem.
2. Never just solve that problem, look for a generalization.
3. Abstracted - looked at the heart of the problem.
4. Search the literature to find out if the problem has been solved by others.
5. If not already solved, solve the abstract problem
6. Solve the original problem.
7. Solve other problems.
8. Publish.

The second part of the talk enumerated a number of problems in today's software engineering research:
- reacting to papers not to real problems, cliques in science;
- confusing: technology, fundamentals, irrelevance;
- last 9 papers scholarship;
- reacting to symptoms rather than cases;
- reacting to funding buzzwords;
- confusing popularity with importance;
- empirical Software Engineering;
- Out of area research (management science, psychology, sociology, economy, abstract mathematics);
- beware of innovation thin line;
- research metrics;
- cooperation with the industry;
- rethinking education in Software Engineering.
While criticizing simplistic experiments and interviews as reliable research methods("people ask people things and they believe them!"), Prof. Parnas mentioned: "That's one of the reasons why I always enjoyed having Liam Bannon around - he's a psychologist! he knows better than that".

He concluded his talk by saying it was "time for a slow careful restart".

Professor Fionn Murtagh presented a briefing entitled “An ICT View of the Current Conjuncture and the Topology of Innovation”, providing some interesting insights into current and future priorities for the SFI in the ICT field.

Update- amateur recordings of the two talks are available here:
And more pictures.

Friday, February 01, 2008

Erich Gamma talk on developing Eclipse and JAZZ

As I am spending time doing fieldwork at IBM's Dublin Software Lab, I had the opportunity to attend a talk by Erich Gamma yesterday. He is one of the authors of the famous computer science book Design Patterns (thus 1/4 of the Gang of Four) and one of the people behind Eclipse and, currently, JAZZ.

In fact, I was in a bit of luck: I first heard of his talk as it was announced that he'd give it as part of the Lero Topics in Software Engineering seminar series at UCD. But that was in the afternoon and I was dreading facing the Dublin traffic to get down there in time. However, by coincidence I overheard some people mentioning the talk and it turns out he was giving the exact same talk at the IBM lab in the morning that same day! No need to travel further from my comfy desk than down to the local auditorium...although I did make a detour to the in-house cafe to acquire a nice, big cup of coffee. ;-)

The title of the talk was
"Developing Software like a band plays Jazz" (see full abstract here). It was a very interesting talk, mainly focusing on the agile practices applied when developing Eclipse and how reflections on this lead them to developing JAZZ and has formed the underpinning values embedded in this software. JAZZ integrates functionality and support for all major areas of software development - development, collaboration, project management, and organisational processes - and mainly emphasises the importance of transparance and general awareness.

I especially liked one of his main points, namely the importance of having an agile rythm in the project. They were working in 6 weeks cycles, delivering a working build at the end of each iteration. This held product quality at a constant high level throughout the project life-time, avoiding the common "hanging rope" trend that many projects experience: the quality and stability of the developed product is at the highest in the beginning and end of the project, but is at a constant low in-between, forming a sort of "hanging rope" curve. With agile 6-weeks cycles, the cruve rather resembles waves, as the quality and stability is raised again at every iteration. Among other things, this is very good for motivation and maintainance of a clear vision of the final product.

It was a very interesting talk that gave an insight into the practices of this high profile software development project (albeit mainly from a process point-of-view, IMHO). In addition, the team I am working with at IBM are, in fact, using a beta version of Rational Team Concert (built on the JAZZ plattform) and are applying many of the same agile principles as the Eclipse/JAZZ team, so most of the developers were attending the talk with me, making it all the more interesting!

Some podcasts about JAZZ and Rational Team Concert:

Kartik Kanakasabesan, product manager for Rational Team Concert, discusses the development environment Rational Team Concert is designed for and looks at some of the challenges Rational Team Concert is intended to address

Lee Nackman, Mike O'Rourke, and John Wiegand, Rational leaders on the open beta of Rational Team Concert Express