In my last Conference Report, I told you about the KS East and West
Meetings in December. March 12-15 found KS Participants back in Atlanta for
the East meeting and, instead of San Francisco, we were in Pasadena this time
for the West.
Requirements figured heavily on the agenda in both locations. At East and
West, Walker Royce, Vice President and General Manager at Rational Software
Corporation and author of Software Project Management, delivered a lively
keynote on the management renaissance taking place within the software industry.
'Paradigm shift' is generally a facile expression to describe any kind of
change taking place these days, but if Royce is reconnoitering the software industry
with 20-20 vision that usage here might be right on target.
|Author and project manager Walker Royce gave presentations in both Atlanta
and Pasadena for participants at Knowledge Sharing East and West.
What Royce sees is an industry, hardly by itself, doing things faster than ever
before, with the traditional approach to the
software development lifecycle process just
no longer able to keep up. Traditionally,
requirements are defined early in the development
process and done with after that.
Royce says this was never really satisfactory
to anyone, but then that's another matter.
What he has observed taking place instead of the traditional approach is an
iterative model, whereby requirements are redefined simultaneously during
design and development. In his book, Royce lays out the foundation of this new
approach by categorizing it into ten principles, the principles of modern software
management, and these are:
- Focus the process on the architecture first
- Attack risks early with an iterative life cycle
- Emphasize component-based development
- Change management of all artifacts
- Simplify change freedom with round-trip engineering
- Use rigorous, model-based design notation
- Instrument the process for objective quality control
- Emphasize demonstration-based assessment
- Plan releases with evolving levels of detail
- Establish a scalable, configurable process
Naturally, we heard a lot about all ten principles at the meetings. Royce is a
commanding speaker, and he brought impressive evidence to back up his arguments
that a change is upon us, and that when applied right will make a difference
for managers who are keen on implementing it.
How applicable is all this to NASA project managers? Many who attended
seemed to think it was very, as they peppered Royce with questions about how to
apply an iterative approach to their own scenarios.
Can it work for you? It's all in the application. Remember principle 10:
Establish a scalable,configurable process. In his book,Royce called this 'Tailoring
the process!' That is, it all depends on the situation. Context is the key.
|ASK Editor-in-Chief lex Laufer and Managing Editor Michelle Collins
Alex Laufer picked up on this theme in his presentation, "Tailoring Project
Processes," delivered in both Atlanta and Pasadena, which included several examples
of how successful project managers from the Navy and Coast Guard, as well
as other federal agencies, practice tailoring. The interesting questions raised by
Laufer's presentation went far beyond tailoring
any one situation to rethinking what
organizations might be like if tailoring were
accepted as the appropriate response to the
amazing variety of situations managers confront
In Atlanta we also heard from Linda
Rosenberg, one of NASA's own software experts, who presented an exercise on
defining requirements. (See Michelle Collins' story in this issue, Lessons From the
Great Masters) The problems with most requirements are mirrored in their documentation,
she argued. At one point Rosenberg displayed a graphic depicting a
bizarre amalgam of a zebra, a cat, and a skunk. "If you don't do requirements
properly, this is what you get," she said, an unrecognizable species of beast you're
not sure whether to offer up to science or straight out euthanize.
|Knowledge Sharing East presenter Linda Rosenberg and Jo Gunderson
enjoy a break between presentations in Atlanta.
Then in Pasadena we heard another take on requirements from John Allmen,
Deputy Division Chief of the Space Projects Division at NASA's Ames Research
Center. Allmen's presentation "Better: the Enemy of Good Enough?" focused
heavily on how to stave off requirements creep. (See how Allmen converted his
presentation into a Practice for this issue of ASK on the following pages) Like
Royce and Rosenberg, Allmen believes clear requirements play a critical role in a
project's success. Allmen would allow mangers room to improve on requirements,
but he would also treat each improvement as a micro project with a separate
budget and schedule of its own. First meet the minimum, then you can
press on. "A requirement is intended to tell you when you've finished your job
on a particular aspect of the project: Good Enough. Go for a better requirement
if you know you can afford to, but remember to maintain sight of the original
requirement and its original budget allocation as you make better."
Overall there was plenty to learn from the four speakers, and from the other
NASA project managers who spoke at the conference and whose stories will
probably be featured in one of the future issues of ASK.
Talk with you again after the Master's Forum over the summer.
|Todd Post is the editor of ASK Magazine and works for Edutech Ltd out of Silver Spring, Maryland. He has written
for many publications in print and online.