''PeterSommerlad wrote this on the XpMailingList:'' I recently read PeterDrucker's "ManagementChallengesForTheTwentyFirstCentury". Drucker explains the concept of knowledge work in contrast to manual labor. He claims that increasing productivity of knowledge workers is one of the major challenges in the 21st century as was the 50fold increase of productivity in manual labor during the last century. Drucker explains FrederickWinslowTaylor's (1856-1915) work in increasing productivity of manual labor. Before Taylor productivity was mainly increased by new tools, methods or technology (Capital), not by real improvement of the productivity of the worker (Labor). I summarize Taylor's improvement approach roughly as: 1. Analyze a tasks constituent motions 1. Record the motions, the physical effort and the time it takes 1. Motions not needed can be eliminated (I like Drucker's "a great many of the traditionally most hallowed procedures turn out to be waste and do not add anything" -- resembles something, doesn't it) 1. Each essential motion is set up to be done in the SIMPLEST WAY... 1. Remaining motions are put in a logical sequence 1. The tools for the task are redesigned according to the new motion sequence (I speculate about some iteration necessary here) (again Drucker says "And whenever we have looked at any job - no matter for how many thousands of years it has been performed - we have found that the traditional tools are totally wrong for the task.") I think Kent and his colleagues did exactly that to come up with some of XP's practices and insights. However, Drucker goes further regarding the current state of knowledge about knowledge worker productivity (KWP). He claims that 6 factors determine KWP: 1. KWP demands the question: "what is the task?" (stories, acceptance tests, feedback, communication) 1. Autonomy: knowledge workers have to manage themselves (separation of technical and business, rights and responsibilities) 1. Continuous innovation has to be part of the work (feedback) 1. Teaching and learning is part of KW responsibility (communication, feedback, pair programming) 1. KWP is not primarily a matter of quantity of output, quality is at least as important 1. KWP requires the knowledge worker to be treated as an asset not cost. It requires knowledge workers WANT to work for the organization. Drucker claims that the first 5 are exactly opposite to what is needed for increasing the productivity of the manual worker. Furthermore, Drucker gives a statement about the need for optimum quality of knowledge work in contrast to good-enough quality required for manual work. Further explanation of knowledge workers tells that they work in a networked fashion and need to be able to understand what information they require from who and what information/results they must deliver to whom. Work on KWP begins with the following questions to the worker: * What is your task? * What should it be? * What should you be expected to contribute? * What hampers you in doing your task and should be eliminated? I'd like to see more mapping of XP practices and values to Drucker's concepts. I think many of the discussions that happened here could benefit from Drucker's insights. His book might also be a good additional reading for doubtful managers or technocrats. Any opinions? ''I agree, there should be some discussion about Drucker in the whole context of XP/agile. I used to think about this a lot, circa 1999, but never really wrote my musings down. You've inspired me to revisit this. --StuCharlton'' P.S.: I liked Drucker's book and recommend it to all knowledge workers and (would be) managers. P.P.S. KnowledgeWorkerBee''''''s don't like hierarchies ... they manage flat organizations with their feet on the ground ... ready to leave when management types create too many barriers to just solving problems. ------ One thing that fascinates me about Drucker is that he has such respect for Taylor -- in fact, he almost always returns to him before discussing ideas of productivity. It's quite fashionable to diss Taylor as a barbarian - Drucker never seems to follow that fashion (nor any fashion, really)... -- StuCharlton See BrainsAsaCheapCommodity