3

Ben Northrop - The Myth of Architect as Chess Master

 3 years ago
source link: http://www.bennorthrop.com/Essays/2020/myth-of-software-architect-as-chess-master.php
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.

The Myth of Architect as Chess Master

(13 comments)
January 6th 2020

For software architects (especially at bigger organizations), it's not that uncommon to get a request like this from a manager or director:

There's this tough problem the developers have been struggling with for weeks (performance, or scalability, or whatever), so can you quickly look at it and show them the best practice, quick fix, etc. that solves it all?

Reasonable ask? No, not really. The expectation is that the architect is a chess master. It's just like that movie we've all seen. There's the young or disheveled or [insert surprising adjective] savant who, walking through central park mid-dialogue with someone else, glances for a nano-second at a pair playing chess and declares "bishop to queen 4, check mate" and then keeps walking. Minds blown.

chess-master.jpg

But why shouldn't an architect be able to size things up as quickly as a chess master? Both an architect and a chess master have accumulated thousands of hours of experience in their disciplines. Both are analyzing current states of a "board" and then evaluating the trade-offs of different next moves or options.

The difference is that chess masters are always playing the exact same game. Rooks always go forward-backward and side-to-side. Bishops always move on diagonals. And because the rules don't ever change, it's easier to take a mental snapshot of a board and instantly pattern match back to some other past experience - e.g. "this is exactly like the game I played/studied four months ago".

Any given software system, on the other hand, may look the same as countless others you've worked on in the past - I mean, hey, it's all ifs and elses, classes and objects, REST and SOAP calls, etc. - but just a little bit below the surface we realize the pieces and the rules are very different. In my system the rook can jump the first 2 squares but in yours the queen has 3 lives. Each system we visit has different intricacies and rules we must learn anew.

This immense variability and complexity is baked into even the simplest software system - it's in the business domain, in the frameworks/techologies used, and in the implementation itself - and so it takes time to grok all these nuances and details before we can build the right mental picture that allows us to reason about it effectively. Even for those with years of experience this is true. Sure there are common patterns, styles, tools, specifications, etc. that, if you know them, can help in terms of more quickly sizing up a system, but almost never do these alone allow you to pull off that type of chess savant "mate in one" mastery. Real-world software solutions take time. Architecture takes time.

This is why I think it's important to think about software architecture as a process and not a destination. Being an architect (or playing the role of an architect) doesn't mean simply divining the "right" solution or recommendation and handing it off to the developers. Anyone can whip up a bunch of boxes and lines that seem plausible, but do they really apply to the game of "chess" you're playing now, or the one from your past project or from that popular blog post? To truly know the right next "move" requires first understanding the rules of the game you're in. What are the business drivers, risks, technical constraints, legacy components, integrations, possible options, trade-offs, and so on. These are the details that take time to identify, clarify, and understand, and they are subtly (and sometimes not so subtly) different from project to project and organization to organization.

In my opinion, the expectation for the architect should be to setup the framework or process by which good architectural "moves" are made (Michael Keeling describes this well in Design It). Thinking an architect is going to stroll into a meeting and find a check mate lurking in plain site is sadly unrealistic.


About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK