Pros and Cons of Waterfall Model
In a minute, we'll take a look at a number of pros and cons of the Waterfall model. But before we do, I first want to cover some of the main high-level advantages and disadvantages to this development process.
The first advantage is that by splitting your project deliveries into different stages, it is easier to maintain control over the development process. This makes it much easier for schedules to be planned out in advance, making the project manager's life much easier. It’s for this reason I've found that experienced project managers tend to favor the Waterfall process. By splitting a project down into the various phases of the Waterfall process, you can easily departmentalize the delivery of your project, meaning that you can assign different roles to different departments and give them a clear list of deliverables and time scales. If any of these departments can't deliver on time for various reasons, it’s easier for a project manager to adjust the overall plan.
Unfortunately, in reality I've seen a plan adjusted where the implementation phase gets squeezed more and more, which means the development team has less time to deliver a working solution. Shortcuts tend to be taken, and the quality can suffer as a result. It's normally code-base unit integration testing that gets affected first. The testing teams in the test phase get a solution that contains more problems, which makes their lives very hard. So while departmentalization is seen as an advantage, it can easily become a disadvantage if another team is late delivering their part of the project.
Now, let's take a look at some of the high-level disadvantages. The Waterfall model doesn't allow any time for reflection or revision to a design. Once the requirements are signed off on, they're not supposed to change. This should mean that the development team has a fixed design that they're going to work towards. In reality, this does not happen, and changes in requirements can often result in chaos as the design documents need updating and re-signing off on by stakeholders.
By the time the development team starts its work, team members are pretty much expected to get it right the first time, and they're not allowed much time to pause for flaws and reflection on the code that they have implemented. By the time you get to the point where you think a change of technical direction is required, it’s normally too late to do anything about it unless you want to affect the delivery dates. This can be quite de-motivating for a development team, as they have to proceed with technical implementations that are full of compromises and technical debts. Once a product has entered the testing stage, change is virtually impossible—whether to the overall design or the actual implementation.
Now we've seen some of the high-level advantages and disadvantages. Let's take a deeper look at more of the benefits of the Waterfall model. Waterfall is a simple process to understand, and on paper it looks like a good idea for running a project. Waterfall is also easier to manage for a project manager, as everything is delivered in stages that can be scheduled and planned in advance. Phases are completed one at a time, where the output from one phase is fed into the input of the next phase. Waterfall generally works well for smaller projects where the risk of changing requirements and scope is lower. Each stage in Waterfall is very clearly defined. This makes it easier to assign clear roles to teams and departments who have to feed into the project. Because each stage is clearly defined, it makes a milestone set up by the project manager easier to understand. If you're working on a stage like Requirements Analysis, you should clearly understand what you need to deliver to the next phase, and by when.
Under Waterfall, the process and results of each stage are well documented. Each stage has clear deliverables that are documented and approved by key project stakeholders. And finally, tasks in a Waterfall project are easy to arrange and plan for a project manager. The Waterfall model fits very neatly into a Gantt chart, so a project manager is generally happiest when they can plan everything out and view a project timeline in an application like Microsoft Project.
The biggest disadvantage of the Waterfall model is you don't get any working software until late in the process. This means that your end users don't get to see their vision come to life until it’s too late to change anything. It can be very hard for non-technical people to be really clear about how they want an application to operate, and it isn't normally until they can visualize an application that they can really give good feedback. You can mitigate this a bit by doing some prototyping in the system design phase to help users visualize their system, but there is nothing like giving them actual working code to try out.
The Waterfall model can introduce a high level of risk and uncertainty for anything but a small project. Just because a set of requirements and a design has been approved does not mean that the requirements won't change. Waterfall is all about getting the requirements, design, and implementation right the first time. This is a grand idea, but in the real world it is very rarely the case, and this is a big risk to a project. We have talked about how Waterfall is better for small projects, but it is possible to have a small, but very complex project. The more complexity that is involved, the more likely it is that change will be needed further down the line. Complexity in the system is also very hard to implement and test, and can often cause delays in the later stages of the Waterfall software development life-cycle.
If you're working on a project where change is expected, then Waterfall is not the right model for you. I've worked on projects for a financial services company where changes in the law were causing compliance regulations to change. Unfortunately, these rules are very open to interpretation, which meant the legal team was involved at a very early stage. This meant that the interpretation changed a few times during the course of the project. If this had been a Waterfall project, we would have been in big trouble, as projects normally come with very hard and fixed set of deadlines.
This project was a perfect fit for an Agile project. If you are working on a large project and the scope changes, the impact can be so expensive and costly that the original business benefit for the project can evaporate, and then the project is cancelled. I've seen this happen a couple of times, and it's a real shame, as projects that show promise are stopped due to restrictions in the process.
Finally, the integration and delivery of a project is done as a "big bang" on a Waterfall project. This means you're introducing huge amounts of change all at once. This can very easily overwhelm testing teams and your operational teams.
No comments:
Post a Comment