Agile Modeling: Enhancing Communication and Understanding
We’ve all seen the statistics, and likely experienced first-hand, the reality of project failure. The majority of software projects continue to be classified as failures. In thinking about this situation we can see that there are a few different ways in which a project can fail (clearly this is not an exhaustive list!):
- the solution does not meet the project requirements (scalability, performance, reliability, cost, etc)
- due to technical challenges we continue to push out deadlines (and in turn, cost) until we reach the point that project sponsors lose faith in the project and cancel funding
- the team does not understand the requirements that have been provided
- the requirements provided are not the correct requirements
If we try to distill this to a high level view, we can see that we have difficulty in understanding the problem and moving to a solution.
However, such a view is too simplistic. We shouldn’t think that the movement from problem to solution is a one-way flow of information - and we definitely shouldn’t see it happening as a one-time, waterfall progression. We’ve seen that before, and it doesn’t go well.
So let’s update our view a bit recognize that there is a flow of information between problem and solution domains and that this is an iterative effort.
Our view of the situation is getting better, but its still too simplistic. Let’s go another layer deeper. We also need to recognize that development is a team sport. We’re going to have multiple stakeholders and multiple people responsible for design, development, deployment and operations.
We’ve taken another step forward in getting a more realistic view. However, there’s one more step that’s worth taking. Although we strive to have our team co-located, the reality is that we work in a distributed, networked world.
So to summarize this progression, project success is limited by our ability to:
- a. Do we understand the problem domain?
- b. Do we understand the solution domain?
- c. Do we understand how to transition between the two domains?
- a. Are stakeholders able to communicate the requirements to those building the solution?
- b. Are those building the solution able to communicate the solution details amongst themselves?
- c. Are the solution builders able to communicate challenges and alternatives to the stakeholders?
And adding to the difficultly is the recognition that we need to deal with proximity and temporal challenges (i.e. locations and time zones).
There are many ways in which we can try to address these challenges - and improve our odds for project success. But where to start? Some basic Agile ideals (manifesto, principles, and some common sense) give us a good starting point.
Rather than throwing money at a bunch of tools (which I’m sure are very impressive!) and trying to adopt a prescriptive and all-encompassing process, let’s take a different approach. Let’s put more value on people, communication, interactions, responding to change and delivering software. How can we best support our people in this approach?
A key technique that is often overlooked, undervalued or misunderstood is the use of modeling - especially as we move to an agile approach. In fighting back against heavy-weight processes and tool-centered development - modeling was guilty by association. So let’s take a few minutes to set the record straight...
A good place to start is to agree on a definition for modeling. Simply put, modeling is the simplification of reality. That’s it, that’s all. It does not imply the use of a specific notation, tool or process. We’re just trying to look at something that is complex, and make some portion of it easier to understand. As they say, sometimes you can’t see the forest for the trees. Unnecessary details make it more difficult to understand the situation. Better to hide those unnecessary details and focus on the important aspects of the situation.
If we can agree on this definition of modeling, we can take things a step further and consider the idea of Agile Modeling. With Agile Modeling we follow an agile approach to using models to help us to understand and communicate. Apologies for the circular definition - but this gives us a good starting point to pose the question “How do we take an agile approach when using models?”
Much like Agile development in general we use a set of values, principles and practices to guide us in being as agile as possible. Some of the highlights of the Agile Modeling approach include:
- Agile Modeling adheres to the Agile Manifesto and supporting principles. As such, it becomes another practice that you can add to your agile toolkit.
- Models support communication and understanding.
- We strive to create simple models using simple tools. Embrace simplicity.
- We know that requirements are going to change and we therefore embrace change as we create models.
- Our focus is on delivering software not models. We use models when and where they add value. If they do not provide value and move us closer to delivering software - then we don’t create a model.
- We keep models for as long as they are needed. If a model has served its purpose and is no longer needed - we throw it away. This allows us to travel light and not get bogged down with busy work.
- We use multiple models. We use models for different perspectives, different levels of abstraction and for different audiences. For each model we create we understand the audience and goal for the model. We do not create models if we don’t have this understanding of its purpose and goals.
- We use a combination of informal and formal models as dictated by situation, audience and goals. For example, a model could be comprised of simple shapes helping to explain a system metaphor OR be a class diagram that uses the UML.
Modeling supports us in communicating and understanding when we create software solutions. As communication and understanding are two of the most critical aspects of delivering software solutions - modeling is a valuable tool that should not be overlooked.
Break free from outdated misperceptions and incorporate modeling into your agile efforts. Agile Modeling adheres to and aligns with Agile values and principles and should be one of the practices within your Agile toolkit. And by adding Agile Modeling to your toolkit - you’ll improve the odds that your project will be a success!
Join us for Part 2 in this series where we take a closer look at the values, principles and practices of Agile Modeling.
AgileModeling.com: The Agile Modeling home page, an excellent and detailed resource created and maintained by Scott Ambler.
Essentials of Agile Modeling - Workshop focused on providing foundational skills for Agile Modeling.
Disciplined Agile Delivery - community website related to supporting a disciplined approach to agile delivery.
About the Author
Lee Ackerman is the VP Products & CTO at The Emphasys Group. Over the years, Lee has investigated and evaluated new ideas, designed and developed solutions and helped others to do the same. Today he focuses on helping organizations improve how they deliver software through automation, reuse and agile best practices.
 This news - provides an example of a recent look at IT project risk and some startling figures on failures.
Why no feedback?
In my opinion, it does because it's about modeling.
Agile takes modeling as bullshit.
Modeling is no code (or, in case modeling is code, it's called programming)
Also, modeling was always taken by the (now prevalent) XP community as "it does work only on paper anyway, why to do it".
There's no trust in the development community that any model other than code would work, and hence, there's no interest in doing it. It produces nothing to be respected, therefore, it's regarded with a "garbage-in-garbage-out" respect, and is one of the first things to go down when we're speaking about waste management.
Yet modeling could be more than that: modeling could be the way to reach from problems to solutions.
It takes care however to only write down what surely works, it takes care to look at problems from complete viewpoints (that is, from the given viewpoint, the model is complete if there cannot be any side effect (not represented property) which would change the model of the system)
But it takes care. It takes wit and wisdom.
A few people do dare to model non-code (like, we expect text input fields to work), but for most people, code is the holy gray, the one which specifies explicitly every single facet of the system, becoming the system itself.
I guess our profession isn't enough for modeling yet.
But you must always take into consideration the audience, level of abstraction and the fact that anything can be a model not just diagrams.