When is a pattern a pattern

I just got out of TSSJS-Europe where I talked about 3 performance tuning/testing mistakes that I constantly run into while out in the field. While I was called them performance anti-patterns in the talk, I also stated true patterns and anti-patterns are not the work of a single individual. In fact they are not even the work of two or three individuals. Patterns and anti-patterns come from a community effort. This is why I’m suspicious of pattern books that sport a single author.

It’s not that I have anything against single author books or that single author books may not have gotten it right. For example, Enterprise Integration Patterns was authored by two people, Gregor Hohpe and Bobby Woolf. However, every pattern in it was vetted by a large community prior to publication. Gregor Hohpe is looking to follow up that book with yet another and to be clear, patterns in that book will also be vetted by the community. It is this lack of community evolvement that bothers me about my declarations made in Barcelona. The problems pointed out are real and they are most likely anti-patterns. However, they have not been vetted by the community and this needs to change or people should be (rightfully so) suspicious of my work.

Over the next couple of blog entries it is my intention to introduce 3 candidates for performance anti-patterns, Shot in the Dark (as named by Dr. Heinz Kabutz), No-stress testing, and DataLite. If there are any others that anyone else would like to throw in then I would be tempted to start building a performance anti-pattern catalog.

tags: java  performance  antipatterns