Book tells of IT managers on a path toward doom
It's not often that I finish reading a book on software development. Usually I barely get through the first few chapters. Sometimes the methodology is too abstract to work in the real world. Or a book is full of great ideas, but I could never sell them to colleagues or management. Or maybe my mind is numb from 25 years of bureaucratic life.
It's not often that I finish reading a book on software development. Usually I barely
get through the first few chapters. Sometimes the methodology is too abstract to work in
the real world. Or a book is full of great ideas, but I could never sell them to
colleagues or management. Or maybe my mind is numb from 25 years of bureaucratic life.
Software guru Ed Yourdon's Death March: The Complete Software Developer's Guide to
Surviving 'Mission Impossible' Projects reads like a Tom Clancy thriller. The book,
written with Paul Becker, is packed with horrific tales of project managers and
programmers careening toward disaster.
Yourdon's chapters are readily digestible over lunch or a commute. Digestible, that is,
unless your stomach turns queasy with painful recollections.
Software developers who are on what Yourdon calls a death march can suffer divorce,
nervous breakdowns, firing and the occasional suicide.
What is a death march? Yourdon--who apologizes to combat veterans for the
title--defines it as a software project with half of the time, money and staff required
for completion. Or one that has requirements twice what the available resources can
handle.
You could apply the term to a project with a less-than-50-percent chance of success.
The longer and larger the effort, the lower the chance for success. Gargantuan efforts
are so prone to failure that even the largest companies can't afford them--leaving the
death march distinction to the federal government.
"If they can put a man on the moon, why can't they ... " refers to the only
potential death march that ended in success. There are no other similar accomplishments.
We certainly can't say, "If they can modernize the tax system ... " or "If
they can modernize the air traffic control system ... "
Today NASA is on a death march, along with every agency. Its mission impossible is
fixing software for the year 2000. Some agencies are replacing the obsolete code--in
record speed, they hope. Others are scrambling to find source code for the executables
that have so faithfully served for decades.
The failure to use source code libraries and configuration management tools has brought
on the vengeance of a death march.
Why do we commit to death marches? As Dilbert creator Scott Adams said, "People
are idiots." Yourdon itemizes common forms of idiocy: politics, naive promises, naive
optimism, bet-the-farm fledgling companies, macho ideology, intense competition,
government regulations and unexpected crises.
These idiocies lead to various death marches: mission-impossible projects to save the
company or the planet, ugly projects that pave the road to success with programmers,
suicide projects that produce only resumes and kamikaze projects for which honor flows
from sacrificing oneself for the doomed cause.
If you are caught in a death march, Yourdon offers a simple word of advice: triage.
When you have only half the necessary resources and time, something must give. Like the
chief resident in a busy emergency room, the project manager must assess each requirement
for its viability and significance.
A smart project manager will make a cold-blooded judgment about each requirement. Some
will get pushed to the side, where they will quietly recover or expire. Others will get
quick team interventions and survive. Truly essential requirements take yelling, pounding,
jabbing and heroic measures to get through development.
Yourdon classifies death march players as owner, customer, shareholder, stakeholder and
champion. He analyzes the interplay of commitment, trade-offs, negotiating tactics and
strategies. He covers recruiting, staffing, personalities, motivations, team building and
working conditions. He describes good-enough software, best and worst practices, and the
daily workload.
His vignettes about the methodology police and the furniture police are hilarious.
In these times of faster, better, cheaper, the National Partnership for Reinventing
Government should give software team leaders and managers copies of Death March,
published by Prentiss Hall Computer Books. More of us are finding ourselves on death
marches without the requisite survival skills. If you want to avoid a death march, reading
this book is a good place to start.
Walter R. Houser, who has more than two decades of experience in federal information
management, is webmaster for a Cabinet agency. His own Web home page is at http://www.cpcug.org/user/houser.
NEXT STORY: NIH opens a bandwidth closet