Group 3 interoperability notes #255
gregtucker
started this conversation in
Standards: Interoperability
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Have run-time appear under interoperability (doc) standards
Distinguish between MODEL COUPLING interop, and MODELS AND OTHER SOFTWARE (e.g., inference)... or not! Are they all variations on the same theme? Or different things?
Code communicates ability to be controlled from outside.
"Offline" vs "online", one-way vs. two-way, need to define. DEFINITIONS matter.
The proverbial future grad students:
understand inputs and outputs
understand scales
understand issues in scaling up/down (avoiding mis-match)
meta-data enriched input and output (what are units) lack of meta-data enriched is a serious problem
grad student receives something containerized, thus avoiding time and suffering with OS, dependencies
a lot of current headaches are over heterogeneous computing environments
well integrated with common libraries, as opposed to out-of-date and/or poorly supported or unusual APIs, not wired to eg a MS product
slow the rate of sw change (maintenance needs), real problem with keeping up to date
standard APIs, standard I/O functions, containerization
most students lose time to getting stuff to work
standard way to have consistent fields of metadata
Human interoperability is also an issue! How we humans interact with others' code!
Beta Was this translation helpful? Give feedback.
All reactions