Will Your Software THRIVE or DIE?

The future state of software projects can be. predicted (with small degrees of uncertainty).

This software is deployed in self-contained DOCKER containers hosted at MolSSI-- contact: Dr. Paul Saxe (but only trained on a few dozen, older projects.Need more. See below.)>

This means we can now find generalizable conclusions across CSSI projects; specifically, we can predict how many CSSI projects will fail to thrive; find and fix projects in trouble and we can stop projects being forgotten by research community.

So new we have new goal: How early, within a standard three year NSF project, can we predict if new software is destined to die, not thrive?

Need new software. THRIVE =

• • ○ find current (and prior) examples of successful computational science software projects;
  • ○ cluster and generalize those projects;
  • ○ map new projects into what is known aboutthose clusters;
  • ○ offer a scorecard on that project, makes suggestions to improve those scores