When exposed to the German education system, you'll pick up the importance of methodology and planning. In the Business Administration classes, we learned how important it is to do your strategic planning (10+ years) and operational (2+ years). OK, in the late 70s the world turned a bit slower. So when introduced to the presessors of RUP (Rational Unified Process), a heavy top-down, big-bang approach seemed very sensible. Well, it worked (somehow) and I thought I just have to improve my planning skills and spend days with Rational Rose trying to get the full round-trip to work.
When starting Compiere in 1999, I know I had to do something else. Especially with Open Source, the traditional Big Bang approaches did not seem to work. So, I thought, I have to start "without a methodology". After a while, we managed to publish stable daily releases and incorporated feedback very quickly. Initially users and customers were concerned about using the daily releases in production, but our track record was persuasive. So, it became standard procedure to use a daily build in production, especially if an important bug fix /improvement was made available.
One day, a friend told me that "our agile approach" was working nicely. I was surprised that what we did had a name - so looked it up and was astonished to find the pricipals I was living actually written down. Happy to have found a name for what I was practicing for quite a while, I signed the Agile Manifesto in early 2002
As agile became more popular over the years, I saw quite a few "agile" projects. The ones which failed usually used "agile" as an excuse for no planning. Not knowing what the requirements (= intended outcome) are or even stressing that they don't want to know - is nice if you are just exploring or playing around, but in my experience the main cause for project failures.
As the founders of Compiere, Kathy and I will now concentrate on the "last mile" - the implementation of Compiere. In our trainings, we stressed the importance of Conference Room Pilots (basically an agile way of doing ERP implementations). To facilitate this process, we are looking for an Agile Project Management tool. There are quite a few good products out there. Our finalists are Rally, Version One and Target Process.
All of them have free Community Editions. We concentrated on the professional editions as we require for the support of multiple projects (= implementations). So, as usual, we came up with our list of requirements and evaluated the products. Well, this is tough - Kathy as the "requirements queen" and implementation project lead spent quite some time evaluating the products. Yes, there are differences, but from a business analyst point of view, you can achieve what we wanted with all products. Some menu systems were more helpful than others, but for the functionality we need they all are all quite suitable. The UI is always important, but also there after getting used to, there are not many differences.
So, seems it comes down to integration (Eclipse, SVN) and ultimately also Compiere. Well, I have not finished that yet - so our decision will be in our next blog.
