This is how software should always be developed – Day 3, Friday

Jesse came in Friday morning and was clearly enjoying our technology stack and working environment. We picked solid and appropriate technologies for the work we were doing and kept management overhead to a minimum. We were really just focusing on writing good code, testing, reviewing, and applying the feedback into the next round of development.

“This is how software should always be developed” – Jesse

Dev-Technology-Agile-Development-1I think anytime you write software utilizing a third party API you will begin to find places where the API and underlying data is insufficient or incomplete. Luckily, we are pretty nimble here and able to make adjustments to our plans as we learn more about the limitations of the data. It appeared that the data was probably better suited for reporting, but we wanted to try to use it a little bit differently. We hoped that with the addition of the RXImage API from the Nation Library of Medicine we could put together a useful and cool prototype.


By Day 3, development was well underway. Co-location and communication—like in most agile projects—are critical to our team being able to crank out quality code in the limited amount of time. Our team self-organized around task prioritization and selection and simply grabbed items on the left side of our Kanban board and worked them until they met our “definition of done” which for backend tasks included unit tests and for frontend tasks included 508 compatibility.

Moving into the weekend we were in pretty good shape. We were almost ready to get some user feedback from our product owner and fearless leader, Michelle Scheuerman. We kept all our work as transparent as possible, which is really allowed us to move forward at lightning pace.

However, on Friday afternoon, we found we had a small issue when I received an email titled Status Check Alarm: “Jenkins – System Status”. Thank you AWS monitoring for the heads up that our Jenkins server died on us. Like we didn’t already have enough to do this week! Luckily AWS makes it easy to spin up a backup.


This post is number 3 in a series of daily posts by Adam D’Angelo about Dev Technology’s 6-day development of the Drug Recall Information Center (DRIC) for the 18F Agile Delivery Services proposal and development challenge. View the entire series here.