Agile Zone is brought to you in partnership with:

Chris Spagnuolo has been working in the Geographic Information Systems (GIS) field for nearly 15 years. If it involves GIS, he's probably done it...everything from field data collection to large scale enterprise GIS deployments. Through this experience, he has reached the conclusion that the most effective way to deliver value is by the implementation of agile practices. He believes strongly in the effectiveness of agile practices and he is leading a new group to evangelize the benefits of agile. Chris is a DZone MVB and is not an employee of DZone and has posted 29 posts at DZone. You can read more from them at their website. View Full User Profile

'Requirements Don't Change, Despite What Engineers Believe'

01.04.2013
| 2959 views |
  • submit to reddit
Alan Cooper is the author of The Inmates are Running the Asylum and About Face. I once attended a talk of his on Post Industrial Management. Mainly, he discussed how to get non-technical people to have success and achieve their goals with software products.  He did make some very interesting and valid points regarding the management and executive view of software development. In particular, he emphasized that organizational structure and technologies that worked in the past are failing today in the post-industrial bit-centric world. I completely agree on this point. Most of what passes for management today is based on industrial management, but the industrial age is over. We need a new wave of executives to find the right organizational cultures and management tactics to support the new post-industrial knowledge workers.

He also hit on one of the topics that I really think is an issue in software management today and that is the difference between effectiveness and efficiency in software development. I have written about this before and feel very passionate about it, and it's worth mentioning again. Cooper provided a quote from Peter Drucker that I think really summarizes this argument very well:
"Effectiveness is more important that efficiency. Business can decline and even fail at the same time that process reform is dramatically improving efficiency by saving money for the company."
I would agree and would argue along with Cooper that ignoring effectiveness to focus on cost reduction is a primary cause of software construction failures.

Cooper also made some great analogies to describe the way software is developed. He doesn't see software development as an industrial activity. He believes it's more like a pre-industrial craft in that things are made one at a time, they're not formulaic, and they are complicated and nuanced. But, he also see's it as a post-industrial craft with more pieces, built by disparate teams, with abstracted notions and massive interaction between the parts. He described the post-industrial craft of software development as existing in a brittle environment with invisible or intangible products, and working in a world of rapidly evolving tools. I agree with most of that, except I'd like to believe that the products we produce aren't invisible to our end users.

I really liked the points discussed above. However, I found the remainder of his talk to be a little off-base and very biased to a waterfall or modified waterfall approach to software development. In addition to promoting a waterfall approach, he made several misstatements about how agile teams work and plan. His first comment was about planning. He advocated that to successfully manage software projects, you must have detailed written plans. He also claimed that contrary to engineers' claims, requirements don't change. He went on to slam agile practices by saying that agile practitioners say that “We shouldn't plan because things change so rapidly”. Anyone who has managed a software project using agile practices knows this is a huge waterfallacy (a fallacy about agile spread around by waterfall advocates). Agile teams do plan at many different levels. In fact, studies show that over the life of a project, agile teams do more planning that traditional development teams…they just do it iteratively.

Where Cooper really began to diverge from an agile approach to software development was in his description of what he called The Software Construction Blueprint. Here are the required elements of Cooper's “blueprint”:

  1. A narrative description of user personas and scenarios
  2. A behavioral overview in narrative form
  3. Detailed form and behavior specifications document
  4. Detailed code examples showing how software will work
  5. Detailed schedule and test plan

According to Cooper, by definition, blueprints are buildable and biddable. If you build the blueprint described above, Cooper wholeheartedly believes that you can give a fixed price bid that you are confident in. Now, I don't know about you, but I've worked on many projects that produced blueprints or whatever you want to call it, but it was heavy up-front requirements analyses. They were not successful! On top of that, our customers are not paying us for all of these narratives and specifications…they're paying us for working software. Now, according to Cooper, an interaction designer builds this entire blueprint and distills everything anyone needs to know about building the software and hands off the design to a production team that actually builds the software. Because the interaction designer does his job and divine's every requirement exactly, there should be no questions or problems negotiating the software development minefield on the part of the production team.

I have several issues with this line of thinking. First and foremost, I believe this sets up serious silos within the development organization. It says that the production team has no input on design and design has no real input on production beyond initial design. I really believe that teams shouldn't have different titles and divided responsibilities. The team is the team, period. Break down these silos and strive for more cross-functional teams. It builds a better understanding of the various project roles and makes for more effective teams.

A side issue I have to throw out there is the fact that Cooper is an interaction designer and really came across as selling his services in the keynote (something I found rather distasteful). Afterward, someone asked Cooper if he or his organization ever builds what they've designed. He answered “No, we don't”. I won't elaborate here, but I'm pretty sure you're thinking what I'm thinking.

The bigger issue with his line of thinking is that all requirements can be defined up front and that they don't change over time. We've all worked software projects before and we all know this isn't true. Customers essentially have some idea of what they want, but as the software is developed and demonstrated, they begin to understand what they really want. In agile, we can address those new ideas or changes without any worry that our big upfront design and planning would have been wasted. But, after asking Mr. Cooper about customer involvement in the production phase of his development model, I completely understood why he believes that requirements don't change. My question was simple, “Where in the production process do you share completed functionality with the customer and how do you manage the change that is inherent when a client begins looking at what has been developed?” Aside from the relative hostility with which he answered this question, it was evident that in his model, there is no communication or collaboration with the customer in the production phase. Apparently, it is possible for an interaction designer to not only define every last requirement up front, but they also have the innate capability to anticipate every change the client would make and addresses them in the blueprint. I guess it would stand to reason that if you ignore your customers and not allow their voice to be heard in the production phase, you'd think requirements don't change too!

What really perturbed me was that Cooper said he doesn't believe that the customer even knows what they want, and that the interaction designer knows better than anyone what they need. I take real umbrage with this assertion. Essentially, his opinion was ignore your customers, you know what they need. This to me is a real problem in the software industry. This kind of thinking is exactly what leads to the 65% of features in most software that are rarely or never used (Check out the Standish Group's CHAOS Report). Having someone of Cooper's stature stand up and advocate this is not going to do anything good for the software development industry. And, considering the already low rate of agile adoption in the GIS development world, the last thing that we needed to hear in a keynote session were these toxic, old-school notions of software development. I've been searching for a term to describe this way of thinking and thanks to the keynote speech at the Dev Summit this week, I think I have one now…The Cooper Syndrome. Maybe at next year's Dev Summit, Dave Bouwman, myself, or someone else can give the agile counterpoint to the Cooper Syndrome.

If you want some more insight on the keynote, check out Jeff Certain's blog post on the keynote.
Published at DZone with permission of Chris Spagnuolo, author and DZone MVB. (source)

(Note: Opinions expressed in this article and its replies are the opinions of their respective authors and not those of DZone, Inc.)

Tags:

Comments

Mark Unknown replied on Fri, 2013/01/04 - 2:50pm

According to the blog, this was "Posted 20th March 2008"

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.