Impressions from ICSE 2010 – part 1

After a long tiring flight to Cape Town (and the loss of my mobile in the back of the taxi), I arrived in Cape Town late Friday (30th) evening. Everything looked very new – in fact my hotel is less than a year old – no doubt due to the World Cup next months. People is very friendly and appear very keen on demonstrating the very best of South Africa, although it does occasionally show a level of inexperience – going along with the newness of everything. Sunday was my first conference day, and it was straight into the ‘fire’ – I was the second presenter of the day.

The ICSE conference (for me) runs over six days broken into the 3 pre-conference days (listed below) and the actual conference:

• Day 1: 5th Workshop on SHAring and Reuse of architectural Knowledge (SHARK)
• Day 2: Tutorial on Cloud Service Engineering
• Day 3: ICSE Doctoral Symposium
• Day 4 – 6: The main conference

Day 1: 5th Workshop on SHAring and Reuse of architectural Knowledge (SHARK)

The workshop consisted of two sessions. The morning session was the presentations of the accepted research and position (opinion) papers, and the afternoon session consisted of three working groups. Each group discussed one of three questions (phrased based on my some what fuzzy memory):

1. What are the core, required architectural knowledge?
2. How can web 2.0 and social computing aid architectural knowledge?
3. How can architectural knowledge repositories best be integrated to address research challenges?

Lot’s of thoughts, ideas and discussions to be summarised and publicised later – and it reaffirmed to me that software architects – academic or industry – can spend hours and hours on discussing concepts, definitions and potential solutions. Software architects from academics really aren’t that different to those working outside research – although they are probably better at asking questions and maybe not as experienced in driving a forced outcome.

Later that evening, the attendees met for further conversations at a local restaurant.

Organisers: Paris Avgeriou, Patricia Lago, Philippe Kruchten

Day 2: Tutorial on Cloud Service Engineering

Stefan Tai appeared as the main presenter – a former researcher at IBM Research, Watson. The day was a world wind tour of Cloud Computing mixed with small programming exercises using the Amazon Cloud. The main points from the day included:

• Cloud Computing is clearly an industry initiative, but the academic world is running very fast to catch. Especially in Europe where a number of major national and research clouds have recently announced.
• Most programming services are still fairly primitive compared with what large Enterprises might expect – e.g., data integrity are now assumed to be managed by the cloud application. And the transparency of the Cloud breaks down in a number of exception handling cases without necessarily providing adequate tools for the application to gracefully address the exception.
• The fundamental architectural assumptions and principles are quite dramatically different to those utilised as part of ‘traditional’ software development. This is a significant difference and potential barrier for adopting among large Enterprises. In the case of Service Oriented Architecture, it was feasible to build a ‘SOA façade’ to existing legacy applications and thereby be able to exploit the benefits of SOA. This doesn’t appear to be a real option in the case of Cloud Computing – an underlying level of re-engineering appear to be necessary in order for Enterprise to ‘Cloud enable’ existing applications.

Day 3: ICSE Doctoral Symposium

Today is a day of workshop with the other Ph.D. students at ICSE – conducted similar to a mini-conference. The day is about the sharing of idea and experiences, but also a day of scary cases of student who are in year 3 without a clue of how to perform evaluation of their research. It seems to be a bit too common case of supervisors not pro-actively bringing enough focus. And also a lack of knowledge of existing research across related areas. Is this a gap in how educational institutions approach to and management of the Ph.D. educational process?