t recent TUG education committee meetings, the members of the group have been choosing speakers and topics. Many of the well known speakers will be returning in 1996 with both ever-popular topics and new topics. Planning for education for six to nine months in the future is an exercise in clairvoyance and to a certain extent we rely on our most well known speakers to be up to date and present us with new topics.
IBM with the AS/400 is certainly making sure that we do not run out of additional "neat" things to get our fingers into. But education is not only about studying what is new and trendy, whether out of interest or duty or both. We need to occasionally take stock of the directions in which we are heading, and for most of us, we need to ensure that we can continue to make a living doing what we enjoy.
This was brought home to me recently when I attended a presentation by Ed Yourdon at a CIPS dinner meeting. Ed was the original guru of structured programming, and in 1991 predicted the decline and fall of the American programmer. In Canada, some of us have experienced the closing of I.S. departments here in favour of consolidation at U.S. head offices. Stories abound of programming moving off-shore. Programming teams can be had in Russia for $200 a month, programmers in Cuba for $20 a month, and the quality is good. Ed, however, is now predicting the rise and resurrection of the North American Programmer.
which can gravitate to wherever it is done best. North Americans should be focusing on new technologies and applications where innovation and creativity are more crucial. Client-server applications and web browsers and builders are examples of areas where North America appears to be ahead. However the concept that made me most sit up and take notice was that of "good enough" software. I quote:
"Zero-defect" software is desirable(!) for nuclear reactors and pacemakers and air traffic control systems... But (it is) unnecessary, impractical, and even undesirable for many other applications. What users really want is software that's cheap enough, fast enough, feature-rich enough, and available soon enough - i.e. "good enough". "
This approach is obviously working in the personal computer world. Witness the compromises made by Microsoft in Windows 95. Is it relevant to AS/400 programmers and systems programmers? PC-based development tools are very attractive and can improve productivity, but can we say goodbye to quality assurance and ISO 9000 registration in favour of timeliness? My recent experience of improving my company's network infrastructure and communications capabilities certainly introduced me to implementing the closest to the users' needs (wants!) that we could achieve, within the time and money constraints. I'm not convinced that a perfect solution to that problem actually exists, and I certainly did not have the time to investigate every possibility.
My philosophy is to identify those people with the appropriate knowledge, especially those who thrive on the satisfaction of getting a job done well, and let them get on with it. But some quality controls are necessary in most circumstances. Must we make a choice between "fast and good enough", and "less timely but error-free"? Or will the newer tools permit us to be inexpensive, fast and error free? Anyone with a proven solution is invited to give me a call.