A brief history of Kilo

[gk-brown]

As with Sierra, elements of Kilo originally began as part of Apache Pivot, a project that is now retired. For example, Kilo's BeanAdapter class is a direct descendant of the Pivot class of the same name. Some of the code in Kilo's JSONEncoder and JSONDecoder classes originated as part of Pivot's JSONSerializer class. ResultSetAdapter was inspired by Pivot's ResultList. WebServiceProxy is conceptually, if not directly, related to Pivot's web query classes.

Kilo was initially called "HTTP-RPC", and was intended to provide a thin, API-centric wrapper around the more general HttpUrlConnection class. A Swift version was also part of the project at the time. When focus eventually shifted to the server, the Swift code was moved to a separate project. The project was renamed in 2022 to reflect its more general intent.

Kilo's primary server-side components, WebService and QueryBuilder, were created as an alternative to other frameworks with a larger footprint and steeper learning curve. WebService has remained largely unchanged throughout its existence, though it has received many updates over the years. QueryBuilder started out as a class called Parameters and supported query construction exclusively via string literals. The current annotation-based implementation is the third iteration. A previous version based on enums proved too cumbersome for practical use and was ultimately abandoned.

As the README notes, the methods of Kilo's Collections and Optionals classes were originally created to provide a more flexible/less verbose alternative to similar functionality provided by the JDK. However, they also represent an attempt to emulate some useful features of other languages, such as the null coalescing operator and null-safe calls, in a Java-friendly way.

Version 5.0 adopted Java 21 as a baseline. The most recent version (5.1.1) was released last month and includes various bug fixes and enhancements.