Today I have started a new project. This looks quite interesting as we will be using the .NET Framework 2.0 (betas), Visual Studio 2005 and related tools. We will also use SCRUM as the project management methodology.

Today we had meetings with the whole team, composed by both consultants (from my company) and customer’s employees. We have then discussed the principles of SCRUM and how we are going to deal with the project.

You know, I have been studying various project management methodologies over the past years, like the RUP (Rational Unified Process), MSF (Microsoft Solution Framework) and XP (eXtreme Programming). They are very different from each other, ranging from more conservative approaches like in RUP to more extreme like in XP.

While I haven’t seen yet a successful implementation of XP in a real life project, I have successfully used both RUP and MSF on my projects and I’m quite confident that they work.

Then we come to SCRUM. It sits in between the previous PM methodologies I listed above. It is not extreme as XP, but it is certainly far away from MSF and RUP.

SCRUM preaches things like no specs and no project manager. It says that the person who does the job is the best person to decide how to do that and the specs are created on demand.

While I partially agree with some of those concepts, I’m really concerned about the architectural quality of the final software if you can’t envision the full final product before committing to a design. It is like building a house without a project. I put up a wall and see if it is good, then I put up another wall and so on. The results are unpredictable.

I have to see in real life, along this project. The project is long, though and will probably keep me on that customer for the next 9 months.

This week I have had the opportunity to meet Bela Ban, the head behind JGroups. For who don’t know, JGroups is a neat piece of technology that provides reliable multicast protocols for Java. It is something new and extraordinarialy useful. JGroups have been used as part of the foundation of JBoss Cache and JBoss Clustering, two pieces of the JBoss J2EE application server.

He is a nice chap and is very sharp. Nice to talk with and certainly he knows lots of stuff and people. He has some ideas about reflection that I don’t agree, but I like him anyway. :-D

The reason of the meeting was to discuss a potential effort to create a JGroups counterpart for .NET. Although .NET has got some very powerful networking technologies implemented in the right way, there is the need for something like JGroups.

I have some ideas about what I would like to see on JGroups .NET or whatever it gets called and it includes native communication with Java, a way to push Java objects into the wire, get them form a .NET application and call methods on them and certainly all the features that a reliable multicast can provide.

We will discuss further this .NET JGroups stuff and see what happens. If my full time job allows me, I will love to get this stuff done.

My new customer is not in London. I will be working 50 miles on South of London in half way to the end of the isle. Today I did this trip for the first time and it is surprisingly fun. From a certain point the motorway becomes a sort of b-road with one lane for each direction and some very beautiful bends. It is the perfect place for some spirited drive and is the best way to start the day before getting to the office.

Well, that’s all folks, see yah!