So How’s It with Architecture - Up-Front Design or Evolutionary Architecture?

Table of Contents

So How’s It with Architecture - Up-Front Design or Evolutionary Architecture?

Where does architecture currently stand? We can say that there are two classic approaches:

  • The classical approach, which requires carefully planning as many details as possible (up-front);
  • The agile approach, which dictates making decisions as late as possible and developing architecture through refactoring.

How does this usually happen in projects? In many cases, a more or less detailed project is created in the up-front style and remains that way. On the other hand, relying solely on organic development of good architecture through natural evolution usually also fails. In large projects, it is extremely risky as it often leads to local solutions that should be rewritten at some stage (which usually does not happen).

In practice, a mixed approach works best. At the beginning of a project, release, or iteration, a concept and design of a solution is created, which serves as the basis and reference point for project work. This project does not have to be, and even should not be, exceedingly detailed. On the other hand, we should not assume that what was devised at the outset will be a perfect solution. Consequently, during the project work, we make ongoing local modifications to the design assumptions through refactoring.

By doing so, we obtain a natural process of architectural development. Initially, it is pre-designed, which prevents us from wasting time and resources on evolutionary wandering. We use evolution to improve the original design. If we link this with the process of Natural Order refactoring, we are in luck! More on that in future posts.

The process sketch looks as follows:

(Sketch image removed)

(Text translated and moved from original old blog automatically by AI. May contain inaccuracies.)

Related Posts

Natural Course of Refactoring online on InfoQ

I am so delighted that the article about Natural Course of Refactoring is live. It is a very simple, yet powerful idea about refactoring (but not easy after all) and I hope this way it will reach more people than ever before. So please retweet it and share it wherever you can.

Read More

Renounce Nothing, Attach to Nothing...

Renounce Nothing, Attach to Nothing…

Today I experienced an “AHA” moment in a certain area. Since the topic is universal, I decided to share it. This is somewhat a continuation of a post from a year ago.

Read More

Building Knowledge in Teams: Main Mistakes and Strategies

Building Knowledge in Teams: Main Mistakes and Strategies

The topic of knowledge management in teams is largely overlooked by IT leaders. There is a silent assumption that it happens automatically. To some extent, it does, as programmers are accustomed to constantly learning to stay ahead. However, this is not sufficient. It’s not enough for everyone (or realistically, half) to learn individually. If the team is to be effective, cohesive and up-to-date knowledge is needed. Knowledge of new technologies or solutions holds limited value compared to this.

Read More