A Consultant's View
Prairie Trail Software, Inc. ............................................................. April 2006
While many organizations look to ISO 9000 compliance, others search for a means to define the free-wheeling ways of most developers.
ISO 9000 is a way of defining everything from a management point of view. Companies that implement ISO 9000 generate lots and lots of documentation.
Unfortunately, such documentation becomes a drag on rapid change. It has to be collected, managed, and updated on a regular basis or it becomes useless. Too often, the documentation codifies something that should have been changed, not codified as the "way things are done."
Mrs. Fields Cookies once had a competitor that operated with lots of documentation and control. The founder of that company talked about how much control he wanted over the specifics of how cookies were made. That competitor is no longer in the retail business.
Mrs. Fields does have strong rules, but the rules are implemented differently. There is strong reporting about what is happening at the stores, analysis of trends, and suggestions about how to improve. The recipes are the same for all the stores, but how to sell them, how to get customers in, and what the customer prefer are not. Mrs. Fields focuses more on communications than on control.
In the supply chain field, some try for extreme control and some try for extreme communication-but then stuff happens to mess up the supply chain. The hurricanes in the gulf not only messed up gasoline production, they also messed up how gasoline was moved from one place to another. Organizations that tried to keep good control (such as FEMA) took time to respond; whereas, organizations that had good communications (such as hospitals) did what ever it took to get the job done.
One of the problems is deciding who to give the visibility to and what visibility to give. Often, documentation is done for the wrong reasons and given to the wrong people. In many industries, one can't spend the time to write things down and run them through committees. When Just In Time manufacturing is running out of supplies, the right people need data on where things are Right Now!
In software development there are two radically different needs: quick availability and high reliability. Banks and financial networks are known for their focus on the latter. Change is very slow for them. Visa may issue new requirements and allow five years for implementation.
In an effort to provide a means to develop software faster, but with higher reliability a number of people are working on "Agile Methods", including the modeling, development, and documentation of the system. While banks might not like Agile Methods, other organizations might like the results.
The point of Agile Documentation is communication. In the software arena, most documentation is overhead-it doesn't help the developers, it doesn't indicate where the software is failing, and it slows the developers down. Under the Agile Documentation rules, documents are created only when they are needed. The documents have a specific audience and are written to communicate specific issues.
The hard part is managing an Agile Development.
Under the "document everything" rule, development was managed by the documents-code was almost secondary.
Under Agile Development, the documents don't really matter, the code does.
However, to know whether or not the code is doing its job:
1. The job desired has to be well defined
2. The code has very specific tests that it has to pass.
All this points to other types of documents that may actually move the process along. Too many times, developers object to documentation because they don't see how the document will help them. By focusing the effort on how to better communicate instead of how to control, the process of documentation is improved.
ISO 9000 might be good for a company like McDonalds where what is served does not change very quickly. In areas that are not as well defined, or those that change rapidly, other methods of managing and documenting are being explored with some interesting results,
Over the years, many an expert has issued his or her forecast. These "experts" range from the local astrologer to high priced consulting firms. While it is humorous to look at the astrologer's predictions one year later to see how many have come true, businesses often make critical decisions based on the consultant's predictions. But the latest research shows that the "experts" are no more likely to be accurate than anyone else.
Why are the "experts" so wrong? Some research shows that "experts" often rely on earlier studies that are faulty. The "experts" often get enamored with the "whiz bang" technology. But mostly, the "experts" are experts in only one aspect of the situation when accurate forecasting needs comprehensive knowledge. Many forecasts neglect one part of reality, either technology, finance, production, resources, or marketing.
For example, the Gartner Group is making money off of their forecasts. A recent forecast was that IT groups would continue to shrink as more and more functions are outsourced to other countries. Yet, they had previously forecasted that nearly all call centers would be outsourced. That is not happening. So, what happened? That forecast looked only at the cost factor and did not consider the marketing backlash that would happen when customers received lessor quality service.
There is one type of forecast that comes closer to reality: one that aggregates a number of individual forecasts. We see this technique in race track betting, the stock markets, and in the Delphi forecasting method. In all of them, individual forecasts are aggregated. In one hurricane center, forecasts are made through this method-each individual forecaster makes a prediction and all the predictions are aggregated.
Web sites are starting to use this technique by opening up a "betting windows" for people to bet on a prediction. The Department of Defense is one such player. The goal was to try to use market forces to get accurate forecasts of events. The problem is that such forecasts can't be as explicit as some people would like (and for the forecasts to be explicit would raise serious moral issues).
What are we left with? The future is always murky. Prophecy is hard to read and interpret properly. The best we can do is to watch the odds and have some insurance if we come up on the wrong side of the bets.
.
Dave Randolph,
President, Prairie Trail Software