Simulation automation options - the magic and fragility of fingers vs scripts.

In this video I look at the strengths and fragilities in the automation of workflows associated with data mining. I also pull together my observations of the folk who create automation agents (my colleagues as well as myself). I think it it time that we revisited our habits and attitudes. Although automators would prefer that simulation tools, their interfaces and file formats are carved in stone so they can get on with their magic without ever worrying about it breaking there are lots of others (like clients and researchers) who want simulation to evolve. So often there is no test that these agents remain fit-for-purpose. I go over coffee break scripts from the old days when we had to do manual transcription of workflows to key capture simplification of those tasks. I also look at the grand edifices which drive complex workflows such as the traverse of a matrix of modes for assessment. Their complexity as well as the terse syntax magic they often contain can make them difficult to maintain or worse, drive them to an early end of life. I discuss a too scary to touch grand Perl edifice and how its replacement ended up embedded within the simulation tool source code. The video ends with pleas - that automators consider how to make their work more robust and maintainable as well as considering the option to ask for tweaks on the simulation side. Tool developers just might say ‘can we get that patch to you tomorrow morning’.