This one sentence explains why agile doesn’t actually work and why Enterprise Architects are still struggling to explain why they are needed.
Because it’s simply not true. Yes, you can change the software more easy in the mind of the software developer, but you can almost never do so in the mind of the user.
I work in a municipality that employs 10.000 people. IT is there to support them, but too many of our systems do the opposite because they were developed by agile companies, who didn’t look at the whole picture and who never understood that the software they sell isn’t the most important thing in the world.
I can already feel the author cringing at the whole picture part, but trust me. I’m Danish we have a nation wide enterprise architecture called “rammearkitekturen” (frame work architecture) that has and is being developed by some of the smartest enterprise architects in our country and what they’ve come up with still can’t contain the complexity of reality, because people aren’t buildings and workflows aren’t actually LEAN and there isn’t a “right” way to do things but instead a myriad of different ways that are all great.
This is where software architecture fails completely and it’s where agile doesn’t understand that it doesn’t actually solve complexity better by taking it in bits. Because real architecture is an architect drawing up a new city hall building that will be capable of housing 2000 employees, a library and city services and then turning those plans over to the engineers who will then build it on time and within budget with minimal changes during the process.
That’s architecture. Enterprise Architecture is mainly about creating a bunch of artefacts that will sit in a virtual drawer for 5 years until the next “architect” decides to waste a bunch of time updating them.
Agile may work at companies like Google, I don’t know, but on non-tech it’s just a bullshit paradigm that doesn’t show better results, but does give middle managers and enterprise architects an excuse as to why something went wrong that doesn’t make them look bad to control groups and actual decision makers.
Sorry, but I just don’t agree with anything in your article.
