Monday, 8 July 2013

TOGAF++ principle 23: the architect doesn’t know best

I’ve commented before on the charmingly Utopian architectural principles that The Open Group Architecture Framework (TOGAF) suggests.

There’s an unwritten central dogma in the TOGAF literature, demonstrated in the attitudes of many architects, that the flow of control is inevitably from the architect to the implementors. Yet in my experience as an architect, the flow only seems to work that way in exceptional circumstances.
More often, I take the business problem to the implementors, who have far greater understanding of the implementation technologies than I have. They tell me what’s possible, and what the options are. Very often, the feasible options are few; often, there is only one. So in many cases, the implementors just tell me how the solution has to be: end of story.

When there’s more than one feasible solution, then as an architect I may, in principle, have some say in which way to go. But even then, there’s usually one solution that the implementors want to implement, and usually for good reasons. To go against the grain of what they want to do is dangerous and often foolish, because a disregarded implementor is an unhappy implementor; and an unhappy implementor is more trouble than a cartload of monkeys.

So while there are occasionally opportunities for the architect to say ‘do it this way, because this way follows the architectural principles that we lofty, floaty, architects have agreed’, those opportunities are far rarer than the dogmatists pretend. I’m led therefore to a new architectural principle:

TOGAF++ principle 23: the architect doesn’t know best
Statement:

The implementor’s advice is worth more than the architect’s principle*.

Rationale:

Implementors understand the detail of the technology far better than the architect ever will, and their goodwill is essential to project success.

Implications:
Solution architectures should in general be based on the implementors’ advice. When implementors of different solution components propose incompatible approaches, other architectural principles may be used to select the better approach. Where an architect has to disregard an implementor’s advice, she or he must take the greatest care to ensure that the implementor, who is a critically important stakeholder, is kept sweet.


* Pedants may wish to add 'except for principle 23'.
(More pedestrian principles will follow in later blogs, if the TOGAF chaps don’t get to me first.)

No comments:

Post a Comment