Chat log from BMI Council meeting

[mdpiper]

There were some important ideas in the chat from today's meeting (where Scott and Joe presented their recent work with BMI).

Here's the chat log, with a little cleanup:

12:33:17 From Allen Lee to Everyone:
Overall I like the approach and the simplicity of this API design +
enhancement - I'd love to see the var_info struct formalized better with
schema validation etc, not just encoded as a C struct. It could be widely
useful for model interoperability and description.. haven't seen much uptake
or use of ontosoft but that seemed promising as well

12:33:54 From Scott D Peckham to Everyone:
https://docs.google.com/document/d/15PVvqOk-9dKZpNfZmzR6uasVYaaPkfyp/edit

12:36:25 From Jess to Everyone:
also fyi, here's a c based, topmodel, implementation of extensions Scott
discussed:
https://github.com/NOAA-OWP/topmodel/blob/owp-bmi-enhance/src/bmi_topmodel.c

13:26:31 From Ryan Cabell (NCAR) to Everyone:
The OpenGL standard has that exact sort of extension-querying API

13:28:23 From Mark Piper to Everyone:
BMI get_version() request: #8

13:30:42 From Niels Drost to Everyone:
Or a single character for every extension? Lots of letters in the alphabet ;-)

13:33:36 From Allen Lee to Everyone:
is the use case for capabilities / versioning obviated by increasing usage
of singularity containers in HTC/HPC where you encapsulate all your
dependencies instead?

13:35:07 From Ryan Cabell (NCAR) to Everyone:
Extensions provide a way to organically grow the standard as well -- if an
extension is so popular that everyone uses it, it's a good candidate for the
next version of the core standard

13:35:51 From Allen Lee to Everyone:
agreed, community supported extension points -> core

13:36:11 From 458 Rich & Connie to Everyone:
I like that idea Ryan

[mdpiper]

An idea, maybe, for the next meeting: we could use Discussions (which is nice) instead of the Zoom chat (which is primitive).