 |
privacy |
||
|
Hangout for experimental confirmation and demonstration of software, computing, and networking. The exercises don't always work out. The professor is a bumbler and the laboratory assistant is a skanky dufus.
Blog Feed Recent Items 
The nfoCentrale Blog Conclave nfoCentrale Associated Sites |
2005-11-16Steve McConnell Beyond Myths of Rapid DevelopmentTen Myths of Rapid Development. This was my second opportunity to hear Steve McConnell speak in the six years I’ve been living in the Seattle area. I enjoy listening to his presentation and the delivery. I recorded the ten myths in my notes in this fashion:
It’s necessary to go to one of these presentations to see how these topics are woven together as each myth is dismissed. McConnell’s perspective is grounded and practical. I have done little justice to the presentation by listing the headings, but it would serve even less to attempt to fill in the blanks. Instead, I want to share some of the impressive ways that Steve promotes professional approaches and attitudes and how that came up in the course of questions and discussion. First, around the cost of defects, this observation passed on by Steve deserves to be carved on the walls of buildings and the sides of mountains as a reminder of the cost of quick-and-dirty: “dirty remains long after quick is forgotten.” In response to a question about teams feeling they must low-ball estimates (for internal IT projects) in order to gain approval, Steve politely and graciously point out that intentionally low-bidding a project completely undermines management’s effort to reconcile costs and priorities and find the best allocation to fit overall business needs. He quietly hinted at the damaging repercussions as the true cost of the project become apparent and the organization suffers the opportunity cost of having missed what might have been more-valuable projects. It is refreshing to see emphasis on the technical staff taking responsibility for empowering management rather than second-guessing and undermining executive decision processes. With regard to advocacy of change of methodologies and processes, Steve also counselled that there are a large number of factors that impact the pace of change in an organization, as well as organizations simply having their own styles. If one does not find satisfaction in working for organizational improvement from within, it would be valuable to find an organization whose practices already fit with a developer’s own preferences and sources of satisfaction in the work. I thought it was also useful to to hear Steve’s cautioning about methodological features that may be seized upon for self-serving purposes and actually misapplied. Sometimes, it occurs to me, adherents of extreme and agile programming methodologies focus on convenience for developers that may not lead to successful results for the clients and users. For example, there was brief discussion about over-generalizing the impossibility of clear-enough requirements in justifying trial-and-error (but “agile”) processes. This concern for working through requirements agreement struck me as a corollary of the importance of removing defects as early as possible: defects in requirements are certainly the most costly to carry through into implementation. The absence of sufficiently-solid requirements is also something that disciplined risk management could spotlight, although this didn’t come up in the talk. Concerning predictability, there were three repeated observations that I want to emphasize together:
Over the course of the talk, Steve reminded us that many of the ideas and approaches were proved out 15-30 years ago. They need to be promoted still because they are available and not being adopted. I take that to mean that the tools are available, it is incorporation in practice that continues to be the challenge. With regard to Steve’s favored methodologies, he suggested (subtlely) that prototypes that have to be thrown-away are a great approach for eliciting requirements, and that iterative methodologies are certainly valuable, with many levels and phases where iteration can be beneficial. Finally, Steve reported that he has returned to work on his book on estimation. We can expect it in the Spring of 2006. {tags: McConnell Rapid Development Software Engineering} Further Observations and Reflections Early this year, I used Construx Estimate while taking a Project Management Course and I find the support for cones of uncertainty much more understandable than relying on the free CoCoMo tools. I would love to see what Steve’s book can provide in my weakest area of project management and project planning. This meeting was my geek respite from 1-1/2 days at a conference of ethnographers, EPIC 2005, where I was definitely a fish out of water. It was also comforting to be reminded how practised presenters are able to bring ideas to the public without the academic weirdness of reading papers aloud that are written in a form only to be read carefully and privately. Standing outside the entrance to the lecture hall, I was facing Steve and wondering whether it was really him — he is younger looking and fitter than I remember him from one of the Seattle University Project Night events when he spoke. Before I got up the nerve to ask if he was himself and introduce myself again, the faculty host walked up and greeted him. I’m sure Steve doesn’t remember me and I’ll have to reconnect by e-mail again. The Seattle Section of the IEEE Computer Society announced this event some time ago. The event was at the joint campus of Cascadia Community College and the University of Washington Bothell location. This gave me an opportunity to wander the shared branch of the University Bookstore (typical for a small campus and not approaching the famous richness of the store serving the main campus), check out the CIS textbooks (geepers, way too pricey: buy them used on amazon.com), and visit the library (nice, but friends of the library at $100 to have borrowing privileges is a bit steep, especially in contrast to something like the ACM and IEEE Computer Society digital libraries). The lecture hall tables that we sat at had power outlets and RJ45 sockets between each pair of positions. It is interesting that a student IEEE groups also supported the Seattle Code Camp last month, hosted at the Devry University Federal Way Campus. One notable difference between the two event locations, along with the smaller size of the Devry location, was the disparity in available public transportation. Devry was served by one bus route with intermittent weekend service and you can’t find out even that much from the web site. The Bothell campus is a very different experience, and provided me an easy and enjoyable public transportation experience, travelling from Microsoft to the Southeast of Bothell and returning through downtown Seattle and home. The IEEE affiliation of student activities here reminds me that there is marginal presence of the Association for Computing Machinery in Seattle (although a couple of SIGs have chapters and there are some student chapters with out-of-date web sites). This has a long history from when, in my student-programmer days, there was an independent Pacific Northwest computing association. It was at one of those conferences, in 1959, that I saw Grace Hopper speak. I also remember chatting with Edmund C. Berkeley, Fred Gruenberger, and Stan Poley at events in Seattle. I’ll save the Grace Hopper stories for another account, but the impression she made was a definite factor in my going to work for Remington Rand Univac in Seattle and then moving on to (newly-named) Sperry Univac in New York City and Blue Bell, Pennsylvania. |
||
|
You are navigating Orcmid's Lair. |
template
created 2004-06-17-20:01 -0700 (pdt)
by orcmid |