Improve

I have ignored the AGILE and CMMI wars for many years and have forced myself into the fray in the past year. If you are taking one side or the other, you have little knowledge of history and probable less history of Systems Engineering.

DR. Win Royce developed the Waterfall model in 1970. It was for Systems Engineering not Software. If you know something of building systems, you will know that all systems have a feedback loop. So did Dr. Royce’s waterfall model. But there was something more. Dr. Royce, in 1970 stated that the strict waterfall model where we asked what the customer wanted and then went off and built it with no more customer interaction was not appropriate. He suggested using “prototypes” to help the customer and the supplier build a mutual understanding of the requirements and then start with the waterfall model, still with feedback loops. The Software Community grabbed hold to the waterfall model and gave it the description and bad rap that one must complete the requirements stage before the design before the ………. In 1986, fourteen (14) years later, Dr. Royce was invited to republish his waterfall article, unedited again. The software community still did not get it.

Later, groups just like the Agile group sprung up to let the software world that this Waterfall stuff was bunk and you should have experts who did pair programming and care about the customer. I am one of the original authors of the CMM and have been actively involved with the CMMI for 20 years. In other words from the beginning. While the folks who created the Agile Manifesto did so in part to create controversy, they have seriously missed the point. Read “Balancing Agility and Discipline” by Barry Boehm and Richard Turner with forewords by Grady Booch, Alistair Cockburn and Arthur Pyster. and Scaling Software Agility by Dean Leffingwell.

Here is my point. Some of the Agile points are well made. Large cooperation’s with Lawyers who were hung up on point-by-point contracts caused the strict following of the Waterfall model and made a mess of things. BUT that was not the fault of the description of the Waterfall Model.

As far as Agile techniques are concerned. Sorry to disappoint all of you do or die fanatics. Prototyping was suggested in 1970. RUP corresponds to CMMI Requirements Development pretty well. There was a 14 month study just on this.

Incremental Development. It is a product lifecycle that has been around for 30 years and is embedded inside of the CMMI. So it goes for Evolutionary Delivery, developed by Tom Gilb a zillion years ago.

So Agile, Evolutionary, Incremental, Prototype, RUP, customer involvement, and others key volatile AGILE words have been around for at least 30 years and are part of the process assets described in OPD of the CMMI. They are techniques that projects must apply as they make BUSINESS SENSE!.

CMMI is process oriented but remember the quality of the process influences the quality of the product. CMMI is about using processes to help build quality products and services that satisfies not only the technical requirements but the customer in the operational environment by the end users that have to use it. CMMI is a spin off of Dr.Demings TQM. Dr. Deming clearly states that process should not be done for process sake but for business. CMMI and process improvement and techniques like Agile and others are about Business.

Please come to the European SEPG and hear my talk titles “The Agile View of the CMMI “and learn even more detail. What makes sense given the constraints your project faces. There are projects that last 12 years that the Waterfall Model fits perfectly. There are 7 days events where the 4 months gathering of requirements is stupid. You only have 7 days. YOU MUST USE YOUR BRAIN AND PUT BUSINESS FIRST!