The NoPipeline approach

For developing the pipeline, we set out with a few goals in mind:

• Build modules where each modules deals with a specific (computational) task
• Make sure those modules can be reused
• Make sure the module functionality can be tested and validated
• Make sure modules have a consistent API, so that a. calling a module is straightforward b. including a module in a pipeline is transparent and seamless

Please note that nothing in these requirements has to do with running a pipeline itself. Rather, we consider this a bottom-up system whereby we first focus on a solid foundation before we actually start to tie things together.

That’s why we call this the NoPipeline approach, similar to NoSQL where ‘No’ does not stand for not, but rather ‘Not Only’. The idea is to focus on the pipeline aspect after the steps are properly defined and tested.