In research labs, there is often a need to customise software at every step in a given bioinformatics workflow, but traditionally it has been difficult to obtain both a high degree of customisability and good performance. Performance-sensitive tools are often highly monolithic, which can make research difficult. We present a novel set of software development principles and a bioinformatics framework, Friedrich, which is currently in early development. Friedrich applications support both early stage experimentation and late stage batch processing, since they simultaneously allow for good performance and a high degree of flexibility and customisability. These benefits are obtained in large part by basing Friedrich on the multiparadigm programming language Scala. We present a case study in the form of a basic genome assembler and its extension with new functionality. Our architecture has the potential to greatly increase the overall productivity of software developers and researchers in bioinformatics.