day 27

artisans of the electrosphere 4: building better camels

"The School of Athens" George Deem, 1987 The first conceptual block that must be brushed away is the wrong-headed idea that software can be manufactured like so much sausage on an assembly line. Even the pyramid-building production line of the Pharaohs was run by artisans. Even the School of Athens - often cited as the original system of mass education - broke down into small groups called classrooms. This leads to the first rule of the electrosphere: SOFTWARE DEVELOPMENT BY COMMITTEE YIELDS A CAMEL. Most failures of software development start by applying Industrial Age principles of manufacturing to what is essentially a creative or artistic process. For example, in an Industrial Age factory, there are too many people, too much planning, and too much control. Sure, this is usually the direct result of size, but why are large teams needed? Large software development teams kill creativity, stump management, and result in late, budget-busting products. Never try to develop good software with a big group of programmers. Is it possible for small teams to develop significant products? Lets look at some examples. Unix had two parents; the MacOS had perhaps 4-5. The Pentium team was small enough to fit into two phone booths. Software development must be done by small groups of intelligent artisans. Corporate ledger-animals hate this rule. They want to speed things up by adding more programmers, but as Frederick Brooks pointed out in "The Mythical Man-Month," crowding people onto a team only makes things worse. Brook's advice is still valid, but often forgotten by the modern manager. The second rule is harsher and more difficult to grasp: NOT EVERYONE CAN BE A PROGRAMMER. In the early 1980s, when it was fashionable to major in Computer Science, university Computer Science departments were overflowing with teenagers who wanted to be just like Bill Gates. A CS degree was their guarantee of riches and bright lights. Unfortunately, while most of them could get into the University, few of them could get into Microsoft. The hiring managers at Microsoft knew better. They knew that only a very small fraction of these graduates had the talent to program. They knew that programming is an art form, not an engineering discipline. "Give us creative speed-coders, and we will know how to manage them," is Microsoft's approach. By requiring daily builds of each product under development, Microsoft keeps their small teams of creative artisans under control. But Microsoft never forgets our second rule. Unless you like junk food, coding until 3am, and have the natural ability lacking in most people, you don't become a programming artisan at Microsoft. But this is not enough. In our final episode, we point the way to the future of software development. Will there still be jobs for programmers in the year 2020?

artisans of the electrosphere: 1 2 3 4 5 Daily Dose Index