BioModels should keep track of the simulation tools/versions that are known to work with each entry

[jonrkarr]

Due to incomplete support for standards by tools and incomplete annotation of their capabilities, an easy way to guarantee BioModels users good recommendations for simulations tools is

1. When each entry is created/updated, try executing the entry with all potentially compatible simulation tools
2. Record the tools/versions which executed the entry successfully
3. Store this information in the BioModels database (list of the latest version of each simulation tool compatible with each entry)
4. Periodically run a cron job to re-execute each BioModels entry with the latest versions of each simulation (and potentially new simulation tools).
   • Record the tools/versions which executed each entry successfully
   • Update this information in the BioModels database
5. For each entry, display links to RunBioSimulations for the known good simulation tools/versions

[luciansmith]

One other wrinkle that I didn't think of this morning: The biomodels curation team periodically re-curates older models, and (more obviously) creates new entries. The final step in curation of any model (old or new) should result in an updated entry in this table.

[jonrkarr]

Good point, this information needs to be recoded separately into the BioModels database for each version of each entry.

If we want to be really careful, BioModels can also record the hash for each Docker image used (version tags are imperfect because we currently allow version tags to be overwritten in case developers need to make a correction; once BioSimulators stabilizes we could disallow this).