Sunday, September 2, 2012

Enterprise Architecting Past, Future or Present

Yesterday I saw a message on Twitter which created an emotional reaction of mine. I tried to answer in 140 letters (the magic of twitter), but failed. I was not able to express what I really felt and did not made my point, which caused (of course) another reaction. The conversation was friendly and nice all the time, but I have an uneasy feeling that I missed my point completly. Very fitting to that problem there was yesterday also a quote:

“If you can't explain it simply, you don't understand it well enough.”,
Albert Einstein (1879 - 1955)

So, where was my problem really located:
Thought for the day: Is #entarch doomed because it wants what's best for the business and everyone else wants what's best for themselves?
By looking at this it is really eye catching for me is that Enterprise Architects are defined altruistic and everyone else egoistic. I personally highly doubt that the job profession has anything to do with being altruistic or egoistic. This is part of the person which fulfills the role, but not in the role itself.

Furthermore I have no evidence that the role Enterprise Architect attracts altruistic people, I instead believe that it attracts egocentric persons. Also the Deliverables of Enterprise Architects in most cases are to my knowledge far away from being the best for business. And that is the second statement which struck me in that one sentence. The superlative best. That would for me translate into: out of all Enterprise Architecture options the ONE best is recommended by an Enterprise Architect.

But my answer wasn't touching my points good enough at that point in time:
I highly doubt that #EntArch knows what's best for anything. Not to be known upfront, only known by looking back!
First sentence fits to my thinking, so point expressed. With the second sentence I tried to express that Enterprise Architects do not know the future, but can only make fact based statements about the past. Both of this needs some explanation. As I have pointed out my definition of Real Enterprise Architecture is kind of a holistic approach, but here I like to focus on the deliverables directly attached to Architecture:

Looking into the future is an interesting (but not always relevant) thing to do. In GLUE it belongs to the Division Destination. As a mathematical exercise the basic assumptions are:
  1. Only one business (reality is that in most companies there is way more than one distinct business)
  2. Only one System of Systems (In a very IT centristic approach this might exist, but taking shadow IT into account it is quite often still a flat out lie. In most companies there is more than one System of Systems, often reflected in Domain Models)
  3. 5 Applications (Amount of known Applications can easily extend 1000, if unknown Applications are also counted it is easily more than 10.000)
  4. 5 Software Elements in each Application (that translates into a VERY simple Application)
  5. 5 Architecture Options to choose from on each GLUE Deck per year (unrealistic low number, I know)
  6. Each of the options on each GLUE Deck is a realistic one.
Given that fact looking just one year ahead to give an architecture recommendation (what architects are typically asked for) there is 5*5*5*1*1 = 125 options to recommendNow in most cases Architecture changes need some time, so lets assume that we want to look 5 years ahead. Then the number is slightly bigger: 125*125*125*125*125 ~ 30 billion (each of the options give another 5 options) options. So who really believes that he is able to look forward, despite the thing which happens the next second. And remember the reality is worse, so the calculation for one year ahead is already way higher amount of realistic options. So in most cases I observe heavy simplification and self fulfilling prophecies by people working hardest towards their favourites and to a certain degree against what they do not like.

Ok, so lets try to look back. By looking at the past there is a high chance that someone asks for why did we select one option, even though there was this other brilliant option (a competition has decided for). Reality is, it was the same amount of options to choose from in the past 5 years than for the future 5 years. 30 billion options from the past, but ending up in a very specific one. Especially if the Enterprise Architect was part of the decision process (via recommendation) I often observe explanation variants which protect why the decision was taken as it was taken. (The only one I find acceptable is: Given the knowledge at hand at that time of decision we decided for the best to our knowledge at that time).

So  how to overcome this? For me it is a lot about living now, not in the future (GLUE Division Destination), because that specific future is most likely not coming true and not in the past (GLUE Division Defence), because it is gone. That puts up a lot of challenging questions when I am asked about a potential future and recommendations and a lot of challenges when I am asked to inspect and check for compliance, but I try my best.

Comments as always more than welcome to improve the ideas behind GLUE.

No comments:

Post a Comment