Kurly's Logistics Web Development Team Unveils Common Library Strategy and First Deliverable

Need for Standardization Amid Rapid Growth

Kurly pioneered the dawn delivery market after launching its service in 2015 and grew rapidly, but this led to a problem where multiple organizations independently developed similar functionality. Technical standardization was needed to resolve the resulting waste of manpower and time, and to reduce context-switching costs.

Choosing Between External and Internal Libraries

External well-known libraries are advantageous for shortening development time, but come with risks of heaviness, security dependencies, and critical updates. In contrast, internal (in-house) libraries are light and flexible to modify, but require continuous maintenance and immediate response from the authors. The logistics web development team established a principle of using external libraries depending on the difficulty of the problem to be solved, and otherwise building internally starting from small units.

Developing the First Package Based on a Monorepo

The team's ultimate goal is integrating the logistics back office and standardizing the domain/frontend stack. To this end, they built react-boilerplate, managed packages in a monorepo structure using lerna, and chose npm as the distribution channel. To ensure code reliability, documentation, code review, and test suites were established as mandatory principles.

Implementation and Results of the add-separator Package

As the first common feature, a number separator addition feature (add-separator) was developed using TDD. By applying interface-based logic instead of regular expressions, they increased extensibility, and as a result recorded an average 20% performance improvement compared to using regular expressions. This package has been released as @kurly-lfd/add-separator and is contributing to productivity improvements.