Avoiding the classic catastrophic computer science failure mode: 2010 acm sigsoft outstanding research award talk

Many computer science research efforts fail. Some of this is inevitable, since research is risky. But sometimes the agenda of a group of researchers fails because there is a part of the problem that everyone agrees is crucial, but that nobody works on. Often this is because there are not enough rewards for working on it; it is hard to publish and/or there is no funding. This is more common than you might think; I call it the classic catastrophic computer science failure mode.

Design Patterns could have fallen into this trap. Our hypothesis was "A standard description of design techniques will enable new designers to become expert faster and will improve the design process for experts." I don't believe we have proven this hypothesis yet, though there is a lot of anecdotal evidence for it. A key step in proving it is developing the standard description of design techniques, and "Design Patterns" is part of that. There were lots of reasons not to write the book. Who were we to try to standardize design? How did we know we had the right patterns? These are unanswerable questions, so we focused on writing the catalog and let other people decide whether it was worthwhile. Other people publicized it for us. Without those people, it would not have made much of an impact.

Other projects have not been so fortunate. I've seen several projects fail because they needed something like a catalog, but nobody worked on developing it. This talk will describe warning signs of the classic catastrophic computer science failure mode and how to avoid it.