Pattern Hype

Every idea needs it's supporters. And support based on feelings and emotion or simply the credibility of the supporter is reasonably called hype. But there is good hype and bad hype. Here are examples of pattern hype. You decide which is which.

KentBeck and I are well practiced at promotion. We took BruceAnderson to dinner after the first ArchitectureHandbookWorkshop and worked on him for hours. Finally he had to say enough, enough, I believe already. -- WardCunningham

It wasn't me, there was a Jewish ventriloquist at the next table! I must admit though that I said something with that meaning. -- BruceAnderson

At the D.C. OOPSLA, I asked everyone in the patterns workshop to write Ask me about patterns on their badge. I know I'm not the only one that did so. Several conference goers told me they had already had interesting conversations with other workshop attendees before me. -- WardCunningham

JimCoplien has argued at PLoP and elsewhere that too much hype leads to PatternSkepticism.

I was at a 2 hour seminar last fall where a number of people, including me, were first introduced to patterns. Some of them were involved in a new project. It would take more fingers than I have (10) to count the number of people who left talking about how patterns were going to be their salvation -- their magic methodology.

Most, if not all, now see them as a useful resource, not as the only "tool" they need. -- BobHanmer

Maybe GettingUpToSpeed was not convincing enough. -- MartineDevos

Computer Aided Pattern Engineering (CAPE -- vs CASE) or should it be called Pattern Aided Software Engineering? I have talked to some people who, in earnest, look forward to the day when we can generate software systems by inputing patterns. There may be merit to some of these efforts, but the idea can backfire and cause people to become skeptical of the patterns (remember when expert systems were all the rage and was AI to many people, or when high level management thought that CASE tools would eliminate the need for programmers). What do you think about PatternAidedSoftwareEngineering?

- ToddCoram

Technology sometimes overtakes yesterday's patterns, so they're tomorrow's automatically generated code. At that point, they cease to be important to me as patterns. The literature becomes less important, because the knowledge can be captured in a tool.

I can think of examples of assembly language patterns I used 25 years ago, which compilers know how to handle today (things like register allocation).

However, I can't name many of these patterns that have become automated. And I think the very best patterns can never be automated, because they call on uniquely human expertise.

My greatest concern isn't automation per se, but the opportunism that comes from automation. That will give rise to an industry of tool vendors and lots of PhD theses based on pattern automation. This all ties into PatternHype: It's easier to make money selling a tool than it is selling good ideas. At worst, these efforts will lead people to do bad things, as is the case with many contemporary CASE tools. At best, they sidestep the moral imperatives of human comfort and aesthetics that are the core of what patterns mean to me.

See my article, "Coding Patterns," C++ Report 8(9), October, 1996.

-- JimCoplien

Part of the bad hype stems from a lot of PatternsMisconceptions (which in turn, lead to a lot of "bad hype" and more PatternSkepticism - its a vicious cycle). I've started a separate page for identifying many of these PatternsMisconceptions. Please contribute to it so we may better help others to understand not only what patterns are, but also what patterns are not.

-- BradAppleton

(your story here)