I am in charge of prepping the Day0 happening in Cork, for the next Release Planning of the
Web21C SDK Team.
A bit of background: Day0 tradition started a year ago (at the Dublin Release Planning) as an opportunity for the whole team members (located between London, Ipswich, Denver in US) to meet and share knowledge. It happens the day before Release Planning - hence the name - and involves technical presentations and discussions on topics selected by a small number of members and presented to the whole team in - typically - two or three streams.
The process is quite interesting as we're organising the event as a mix of
Open Space and
Unconference. I have asked people to sign up for sessions to present and for topics to discuss at the
Gold Fish Bowl at the end of the day. (Yeah, I know that a preset agenda is not truly Open Space, but it'll help me to manage the hotel resources and keep people focussed, whilst having the attendee to comment and provide feedback to the presenters - and whoever has something to say on the day can book his session at the start of the day on a properly prepared sheet)
I am confident that the event will be as successful as the one I organised in Glasgow last January.
I am even attempting to present two sessions although I have the feeling it's going to be too much, being now involved on preparing the two papers for
Agile 2008 the release of the Messaging capability at the back of the
Web21C SDK next 28th of April and the Brown Bag at the end of May.
Anyway, one of the sessions I was thinking to present is on software metrics and refactoring complexity out of code. The goods the bads and the evils of metrics and how they work as triggering alarms on spotting complexity in the design and implementation. I am thinking how to best make use of those numbers without being too mental on making them nice and round. Especially the
Cyclomatic Complexity, which I find quite useful as an indication of the complexity of the code. In fact with
Eclipse and the
Checkstyle plugin, it's possible to get the number on the fly. Another useful application of the McCabe number (the other alias of the CC) is to get an indication on how many tests are required to fully cover untested code - useful clearly when you inherit untested code.
I'll think of writing a more extensive post once I finish the presentation