Will OpenSees Ever Be Un-Seen?

OpenSees contains sophisticated solution algorithms and element and constitutive models for simulating the nonlinear response of structural and geotechnical systems to natural and human-made hazards.

But sometimes these sophisticated models are put to less than sophisticated use. I’ve been kicking this idea around in my head for a while, but what xykademiqz wrote in a recent post perfectly explains the sentiment:

… we increasingly see very sophisticated simulation tools used as blunt instruments, simulating everything but explaining nothing.

Simulating Everything but Explaining Nothing, or “Seen” if I may, using the same capitalization as “Sees” in OpenSees. The past participle to the present tense!

If you saw the Google Scholar alerts I receive every few days for new citations to OpenSees, you’d know what Seen is all about. Fragility functions for structural systems that will never be built. Optimal designs that will never be constructible. Parametric studies on properties that do not matter. High performance computing on bloated models. Machine learning algorithms that produce obvious results. The list goes on, populated with too many names to name.

And the writing that accompanies those outcomes explains nothing about either the simulations or the nothing. Nobody wants to read that sh!t.

Don’t get me wrong though. Not all of OpenSees is Seen.

But, what can we do about the Seen? Can those layers of OpenSees ever be un-Seen, i.e., generating useful and explainable insights instead of the unexplainable everything? What do you think?