Nowadays, most companies claim that they are Agile, as it has become kind of a standard. The majority of them also say that they build working software, but, in reality, each company understands this requirement differently. And herein lies the problem! At Apiumhub, we have delivered applications for clients from different industries using Agile methodologies, and have come to the conclusion that to be truly Agile we should do TDD and CI, focusing on working software. And when I say working software, I'm not only talking about software from a functionality point of view, but taking performance, scalability, and cost-effectiveness into account as well.
Working Software Formula
As most of you already know, Agile development is all about iterative and incremental evolution. Agile helps software development teams in a changing environment focus on the rapid delivery of value and on building working software. Agile development is based on rapid feedback, continuous improvement, flexibility, and the delivery of results with high quality and great functionality. In the software world, when we talk about Agile methodology, we are also talking about unit testing, Test-Driven Development, and Continuous Integration.1
For example, using TDD, software developers create unit tests when they program their code so that each unit test examines a small piece of code to make sure that these units perform as expected. Also, TDD allows teams to delete, add, or edit functionalities in a fast and very effective way. But let's face it, mistakes will happen. If you want to discover bugs at the initial stage of development and avoid finding them in the build stage, TDD is really a must.
In TDD, you begin by writing tests before coding new functionalities. And you write the production code necessary to make the test pass, and only after that do you refactor the code to make it more maintainable. This method guarantees immediate feedback after changes are made, so, in other words, it makes it very easy and very fast to find new errors. This immediate feedback allows software developers to fix the defects much faster as compared to other methods where code is tested days after implementation. TDD shows you if the code does what it is expected to do in real time. Basically, it allows you to save time and money as well as fight for the first mover advantage because you launch high-quality functionalities faster, which is really key to success in this rapidly developing world of technology. What is more, Test-Driven Development implementation is more likely to match the Product Owner's vision when we are talking about user stories. Tests are easily developed from the acceptance criteria. TDD guarantees that the final version will meet the needs of the Product Owner. Also, what we really like at Apiumhub is that TDD prevents us from having unwanted design or components in the product. It defines the exact set of required features. It makes it very easy to find redundant code, and detect and eliminate unnecessary development tasks.
TDD really goes together well with Agile. It simplifies and speeds up the process of developing working software, making it possible to launch a new, scalable product as fast as possible with a solid and working MVP.
As for Continuous Integration, we think it is a crucial part of Agile software development. It helps developers build working software because it organizes development into functional user stories, which are put into Sprints. When software developers make a new addition to their code, a test for this code can be added to the suite of tests that are already being run when the developers integrate their work. It allows them to make sure that new additions don’t break the functioning work of existing elements, and developers whose code does in fact “break the build” can be notified quickly. Also, I should say that we are big fans of Martin Fowler, we always read his blog and strongly support his point of view. He says that Continuous Integration is a software development practice that requires team members to integrate code into a shared repository, frequently. The idea is that every person integrates at least daily, which leads to multiple integrations each day. Moreover, normally, several people or even multiple teams work on the same project and Continuous Integration helps to prevent integration problems. These integrations are verified by an automated build that runs regression tests to detect integration errors as quickly as possible. The CI rule states that programmers never leave anything unintegrated at the end of the day. So, developers always know what other team members are working on, how they did it, and they are sure that their code works and is integrated correctly; and even if something doesn’t integrate or something doesn’t work, it’s easy to solve. Integrating code frequently leads to reduced risk of failure, i.e. reduced risk of overspending money on redoing work. Normally, once teams have implemented Continuous Integration, they never switch back, they stick to CI, because they find that it leads to fewer integration issues and enables the development of working software more rapidly.
On the way to production, there will be struggles and issues. But we think that a good way to measure progress is “Working software.” What do I mean by that? Working software is tested software that delivers value to the end-user, i.e. a product that works well, maybe even better than expected, but never worse! In other words, working software is a piece of software which is fully integrated, tested, and ready to be shipped to customers or deployed into production. The only thing that counts is completed, fully functional pieces of the wanted product – working software.
Working Software Rules
To build working software, teams must follow several rules. For example:
- If you are not getting everything you planned to finish, then take on less, but do it well. Don’t try to develop all the functionalities, when half of them won’t work as expected. Do half, but do them right.
- If you see defects in your code, test it sooner and better.
- If your code is hard to work with, then refactor it as you go.
- If it is hard to integrate, then integrate smaller pieces more frequently.
- If you deliver your work to your customers and they don’t like it, then talk with them more before you start, as well as during the development process.
So, what can help you build working software? Your commitment and focus, putting quality into your work, active user involvement, a good understanding of your Definition of Done, two-week Sprints, TDD, CI, acceptance tests, and thinking about performance, scalability, and cost reduction.
And once you have built working software, make sure you go live right.
Go Live Strategy and Checklist
- Make sure all unit and functional tests are green. If they do not pass, the artifact is not built properly.
- Once we have the artifact, ensure that all the acceptance criteria are met. Here it is very important to consider user story format.
- If all the acceptance criteria are OK, test the main functionalities of the app manually.
- The logical user flows are covered by automatic functional tests. But, there is always room to make some visual mistakes or errors, so manual QA is inevitable.
If you are interested in how other CTOs go live, check this article out, where 10+ tech startup leaders reveal how they prioritize software development.
I hope this article was useful for you!