Software systems that operate in an international environment often must support multilingual data models. For example, users of a procurement system must be able to describe the products they want to buy in many languages because they want to receive offers from suppliers that reside in different countries. Designing a system that can effectively display data in a given users language and also allow him to do full text searches is a challenge - many commonly used patterns will have a high performance penalty and slow down your system. In this first of a series of articles we will describe how Postgres supports full text search in general and what are the most common anti-patterns for multilingual SQL models.

The times when Java was available on almost every desktop are long gone but you can still distribute your desktop applications written in Java in a user friendly way. Since the release of OpenJDK and OpenJFX 9 we can benefit from the JVMs modularization efforts and easily build executables with a bundled JVM trimmed to the needs of our application. In this article we will show you how we ported our small desktop app called PDF Decorator initially to OpenJDK 11 and used tools like jlink and jpackage to start distributing our app without requiring any third party software on our clients machines.

Problem: we want queries to our entities to eagerly fetch only the fields that we need in the given context (for example to show in a specific UI data table). Requirement: our solution must be able to accept dynamic filter compositions. Possible solutions: Named Entity Graphs from the JPA standard or Projections mechanism from Spring Data. Lets research them!