docs: add glossary, terms section

[cmoussa1]

Mentioned in #185, it might be a good idea to add a Glossary page to our docs site to go into slightly more detail with some of the Flux-specific concepts we introduce in our guides. This could either be a separate page, hyperlinks to footnotes on the bottom of the page they are introduced in, or both.

Definitions of these terms should probably be kept brief; if they are complex terms or concepts, perhaps there should be consideration of making them their own page on the docs site.

I think it might be a good idea to collaborate some on coming up with these definitions (since I am not the one who came up with a number of these concepts and am probably bound to get them wrong 😆), so if others are okay with it, this could make for a good coffee hour exercise! I would be happy to work on adding these to the docs site, however.

[garlick]

A few quick ideas

flux instance: a group of flux-broker processes which are launched via any parallel launch utility that supports PMI (process manager interface)
kvs: a component of a Flux instance where information pertaining to the instance, such as resource information, is stored in a key-value format
jobspec: a file which defines the resources, tasks, and other attributes of a single program
PMI: a standard API and wire protocol for communication between MPI runtimes and resource managers
jobtap plugin: a built-in or external plugin that can be loaded into the job manager broker module
flux-broker: a distributed message broker daemon that provides communications services within a Flux instance
hostlist: a format supported by existing HPC tools to create a compact list of hosts by name, consisting of a common prefix, a number, and an optional suffix
TOML (Tom's Obvious Minimal Language): a configuration file format that can be parsed
policy: attributes of a Fux queue that define behavior of the queue, such as time or resource limits
limits: maximum values that are enforced across jobs, queues, and users, such as time and/or resource limits

• system instance
• single user mode
• job eventlog
• resource definition (R)
• workflow
• high throughput computing
• fractal scheduling (scheduler nesting)
• flux-imp
• flux-shell
• job request frobnicator
• job request validaton
• feasability
• idset
• flux standard duration
• FLUID
• perilog
• queue terms (start/stop/enable/disable)
• drain
• overlay terms (full, partial, degraded, lost, offline)

EDIT: If it's okay, I'm going to edit this comment with definitions I've found by looking through our docs site or throughout our repos.

[chu11]

i'll just throw this out there, would a similar "acronyms" page be useful? or is a glossary enough.

tbon
kvs
imp
fluid
R
edIt: fsd

ehhh the list is sorta short, perhaps longer out if we end up with enough of them

[cmoussa1]

I think that's a great idea. If we don't dedicate an entire page, we could at least spell out the acronym in parentheses the first time it is used in our docs. This is at least the format I follow when I write papers.

For example, the Flux KVS section already does something like this:

The key-value store (kvs)

Although the method I mentioned above reverses this:

- The key-value store (kvs)
+ The KVS (key-value store)

[garlick]

IMHO throw everything in the glossary, acronym or backronym or no.