Some researchers have suggested that a pattern is “a solution to a problem in a context,” citing Chris Alexander’s work in architecture (see the bibliography). Here are some thoughts on this.
- A pattern is a template, not a specific solution.
- Alexander’s “pattern” theory remains unaccepted by his peers.
- No dictionary supports his definition of the word “pattern.” (Although it is true that one may assign any meaning to any word, when a word’s meaning has been established over several millennia of human experience, it is probably unwise to do so.)
- Although “a solution to a problem in a context” is a compelling writing style–after all, nearly every sales letter follows it–that does not make an instance of that writing style a “pattern.”
With the publication of Design Patterns by the “Gang of Four,” Christopher Alexander’s pattern language format caught fire in software engineering. Three decades later, industry authors still publish book after book after book enumerating pattern languages. Even outside of software, the Alexandrian form continues to spread. From pedagogy to presentations, from activism even back to architecture, pattern languages continue to be mined and shared.
Given how pervasive pattern languages have become, I was surprised to see such stinging criticism of it by Peter Coad, author of Object Models, Second Edition. I don’t know whether he wrote this to set apart the 31 patterns in his book, or to prevent readers from sending pedantic letters about what a pattern “should be.”
What do you think of Alexander’s “pattern” theory and its continued use in our industry? Let me know in the comments.
By the way, the Object Models bibliography takes yet another dig at Alexander’s work:
Alexander, Chris. The Timeless Way of Building: A Pattern Language. Oxford University Press, 1979. An oft-cited book, one that focuses on cataloging “solutions to a problem in a context”–something Alexander calls a “pattern” (although no dictionary defines the term that way).
Be First to Comment
Reposts