Upload
ljmcneill33
View
34
Download
6
Embed Size (px)
Citation preview
THE SINKING OF THE VASATAKEN FROM:RICHARD E. FAIRLEY, MARY JANE WILLSHIRE, IEEE SOFTWARE JOURNAL, MARCH 2003,‘WHY THE VASA SANK:10 PROBLEMS AND SOME ANTIDOTES FOR SOFTWARE PROJECTS’
The Vasa, salvaged and restored in Stockholm
SINKING OF THE VASA Aug 10, 1628, Sweden’s most spectacular
ship ever, the Vasa, departed on its maiden voyage and capsized a mile from shore, killing 53 sailors
Why?
FACTS: King Gustav commissioned the
building of 4 ships in 1625:2 smaller ships, 108 ft keel2 larger ships, 135 foot keel
FACTS Ship to be based on
existing nautical technology proven production methods
To be built quickly
Construction started in early 1626 on the Vasa, as a small, traditional ship
Construction ended in late1628, with the Vasa being delivered as a large, ‘innovative’ ship
FACTS: Confusing changes to the requirements
were introduced:2 smaller ships now to be 120 ft keel (from
108), to carry more armaments (war w/ Denmark)
Then, Vasa was ordered to have 135 foot keel, to host 2 gun decks Problem: 111 foot keel already laid No expertise in Sweden for 2 gun decks
Builders decided to extend 111 ft keel, rather than lay new 135 ft keel
FACTS: Confusing changes to the requirements were
introduced: Vasa built 1.5’ wider than originally planned to accommodate
extra gun deck, but only in upper part of ship. This raised the center of gravity considerably.
Presence of 2 gun decks did not provide enough room for ballast
VASA CROSS-SECTION
FACT – ORIGINAL ARMAMENTURE INCLUDED THIRTY-TWO “24 POUNDERS”
36 ‘24 pounders’ 24 ‘12 pounders’ 8 ‘48 pounders’ 10 other smaller
guns
64 ‘24 pounders’
New Specs Final Specs
Final Implementati
on: 48 ‘24 Pounders’
VASA - STERN
FACTS: Henrik Hyberttson, chief shipbuilder,
became ill and died in 1627, midway through constructionHe took most of the tacit knowledge of
the ‘big picture’ with him to the grave, leaving no documentation.
Project planning and communication was weak, lacking detailed specs, milestones, workplans.
400 people in 5 different groups performed uncoordinated work
FACTS: There were no techniques for calculating center of
gravity and other aspects of performance and stability. Captains had to ‘feel’ the ship and make intuitive
adjustments.
Pre-launch stability test consisted of 30 men running across the deck from port to starboard.
Modern simulations have shown that, because of its haphazard construction, the Vasa would capsize at a heeling angle of only 10 degrees, and a wind of 4 knots. *
*4 knots is not enough wind to unfurl a flag on a flag post. It basically means that you can feel wind on exposed skin.
FACTS: The test was halted due to the violent
rocking of the ship. Not much room for ballast below deck Shipbuilder was unaware of the results of
the test; ‘If only the King were here’… The King demanded that the ship set sail
by July 25, lest the parties responsible be subject to disgrace.
FACT The Vasa Sank.
PROBLEMS/ANTIDOTES
Excessive schedule pressure Objective estimates More resources Better resources Prioritized requirements Phased releases
Changing needs Iterative development Change control/baseline management
PROBLEMS/ANTIDOTES
Lack of technical specifications Development of initial specifications Event-driven updating of specifications Accountability
Lack of a documented project plan Periodic and event-driven updating Baseline management of the project plan Accountability (designated project manager)
PROBLEMS/ANTIDOTESExcessive and secondary innovations
Baseline control Impact analysis Continuous risk management A designated software architect
Requirements creep Initial requirements baseline Baseline management Risk management A designated software architect
PROBLEMS & ANTIDOTES Lack of scientific methods
Prototyping Incremental development Technical performance measurement
Ignoring the obvious Back-of-the-envelope calculations Assimilation of lessons learned
Unethical behavior Ethical work environments and work cultures Personal adherence to a code of ethics
THE END(Or “Slutet” in Swedish)