Skip to main content

Fewer Gurus Needed

Posted by daniel on August 12, 2003 at 6:59 AM PDT

Look around at the tasks you perform every day. How did you learn them? When you are trying to lend credibility to an approach you took, do you walk through the code and show the advantages or do you quote a guru who recommends taking that approach. Sometimes knowing who recommends an approach is helpful. It provides context for the recommendation.

In his keynote address yesterday at XP Universe, Ivar Jacobson noted that "the fact that we still need gurus is a sign of immaturity." Using this measure, the XP world is beginning to mature. There are a lot fewer references to what Kent says and a lot more practitioners comparing concrete experiences.

Aside: Don't write me to complain that in the last paragraph I depended on a guru to give weight to my argument not to depend on gurus. I'm crediting the speaker who spawned this thread.

Jacobson has been looking at process for a long time. He thinks that more than five people working together is a large project. More than ten is too large and might require a system of systems. He is concerned that software development is more complex than it has ever been and that it's too hard for developers to keep up. You need to read thirty or more books to learn what you need. By the time you learn what you think you need, it has changed.

The thirty books may overlap, conflict, or have different focus. A book that focuses on a particular technique or technology will tend to see that subject as providing the answers that you need. Even books on the same topic may give contradictory advice. Jacobson says that "people buy books but they don't read them so you have to travel around the world to tell what's in them." The approach Jacobson is now working on is to have separable but composable practices. This way people can work in more clearly defined areas that interoperate well.

Today's featured Weblogs include Cliff Schmidt's first post What do Customers Mean by "Standards". Cliff has been thinking a lot about standards lately and will flesh out this first post, but he wants to consider what standards mean for a variety of possible customers. Greg Vaughn responds to Matt Stephens' Case Against Extreme Programming in his blog entry Fundamental Problem with Extreme Programming. Greg doesn't buy into all of XP and yet he explains that he has benefited by some of the Agile processes.

Joshua Marinacci issues a challenge in Where are my free JavaBeans. He wants "A Freshmeat for smaller bits. The download.com of components." Can we save the small reusable bits of Java and serve them up in a useful way? How could java.net help?

In the Also Today section, we link to two articles on programming with interfaces. Alan Holub's Why extends is evil looks at the benefits of using interfaces over extending base classes. There is a fundamental architectural choice you make when you design a new package or API. Often the ratio of factory methods to classes in a distribution is a giveaway of which choice was made. The issue isn't new. Bill Venners detailed some of the benefits of Designing with Interfaces almost five years ago. The examples in the two articles complement each other nicely and combine to summarize some of the benefits of considering using more interfaces.

From the Java Today News Page, news editor Steve Mallett, has gathered the following News Headlines .

This blog is delivered weekdays as the Java Today RSS feed. Once this page is no longer featured as the front page of href="http://today.java.net"> Java Today it will be archived at href="http://today.java.net/today/archive/index_08122003.html">
http://today.java.net/today/archive/index_08122003.html. You can
access other past issues by changing the address appropriately.