[epf-dev] Re: Architectural mechanisms - a bit confusing

[Christoph Steindl <steindl@xxxxxxxxxxxx>]

I like the way Peter put it and want to give a short reply:

It may really well be that the architectural mechanisms are very 
important to understand and to use.

For the process design decision, let's consider the target group for 
BUP: How big will the typical team be? How complex will the typical 
problem be?

If we conclude that the team is considerably large and the problem is 
considerably complex, that architectural work has more value than if the 
team is small and the problem is not really complex.

If we decide for architectural mechanisms, then please let us describe 
them in a way that is impossible to misunderstand. It should be very 
clear for the practitioner why to use them and how to apply them.

My gut feeling is that BUP could live without them (while it may really 
be essential for RUP), but I'm really happy to be convinced of the opposite.

   Christoph Steindl

> I agree with everything Sigurd said about changing the BUP content.  I 
> also agree with Christoph that creating architectural mechanisms is just 
> one way of getting to the goal.  However, as I understand doing essential 
> and proper architecture work is a key differentiator of BUP against other 
> code-centric approaches.  So the process design decision we need to make 
> is if using mechanisms is just one technique out of many that we capture 
> as guidance or if architecture mechanisms are really driving the process 
> of architecture work and we model them as tasks and artifacts.  Based on 
> my RUP-biased experience I think the latter would be the way to go if we 
> want to make the case to the agile world that this would lead to better 
> results than evolving architecture in the code.
>
> Thanks and best regards,
> Peter Haumer
>