WICSA 2005:Why Design Decisions Transcend a Particular Architecture
From WICSA Conference Wiki
Architecture decisions are about trade-offs. To capture the trade-offs between different architectures requires a description of the alternative architectures. Frequently the alternatives are significantly different in structure, interfaces, and other properties. So the challenge is clear. How do you attach these alternative structure descriptions to the finally realized architecture when they may have radically different structure?
I submit the answer is you can't or don't. The trade-off analysis and descriptions must exist outside the final architecture because it needs to utilize and reference the alternative architectures.
I would like to make two remarks here:
- If you cannot express an alternative structure, how can you communicate it in the first place? So if you can communicate it, you should be able to describe in *some* way. This doesn't have to be perfect, precise, etc. as long it conveys the basic idea of the alternative
- You are talking about a "final" architecture. I believe, no such thing exists in the first place. Architectures live and change over time during the whole of the live-cycle of a system.
Furthermore, what do you mean with outside the architecture? Do you mean the architecturel description?
