Add 'theme' tags to initiatives

[yuvipanda]

We learnt in our interactions with various community members that the large number of initiatives can be hard to browse. There is many there that are of deep interest to them, and many that are not. It's currently hard to find only the things they are looking for.

We can add a concept of theme to help group them so it is easier for people to scan to find things that are of interest to them, and ignore those that aren't.

Definition of Done

• We have a clear definition of what a 'theme' is (@aprilmj proposes defining a theme as a set of common problems shared by a a number of communities, where the theme is made up of multiple initiatives on our roadmap)
• We have 3, maybe 4 themes defined, with an understanding of what fits in there
• We display these in the public roadmap page

[yuvipanda]

@Gman0909 took a first stab at these, with the following themes;

• cost-mitigation
• knowledge-sharing
• multi-cloud-support
• 3rd-party-integration
• resource-management
• access-control
• platform-resilience

[yuvipanda]

I want to focus on themes as ways for people to discover initiatives that are relevant to them. So in terms of looking at the subheadings, themes are focused towards both the problem statement and the proposed solution headings (which both have outward audiences) rather than the proposed implementation heading (which is an inward audience).

Based on this, I'd propose the following tags to account for the majority of our initiatives:

1. control-cloud-costs (Controlling cloud costs, in both research and educational use cases)
2. improve-learner-experience (Improve experience for students and other learners directly)
3. improve-gpu-experience (Improve specifically experience of using GPUs, for both students & researchers)
4. broaden-access (Broaden who can access hubs we run and find use in them, by serving populations who aren't currently served) (could be better named)

We can also define a theme tag as requiring a verb-noun format, where we are trying to do x for y. This helps keep it sharp, as well as make it easier for people browsing to know what is likely to apply to them vs not.

This is only for platform initiatives, not foundational initiatives. I'll work on a different system for those.

[aprilmj]

i like the labels that lean more towards a community's view of the roadmap, so that they can find the things they care most about more easily

i've heard something like "knowledge sharing" or "data sharing" from a few communities more directly (Reflective most recently) which could mean:

• there are less-known practices that we should help people discover, not new features
• we're missing initiatives that address people's sharing issues
• some of our initiatives address sharing issues AND one of these other things (but they'd still be particularly interesting to people whose goals revolve around sharing)

in the verb-noun language that seems to be "share inputs / data" and "share research or class outcomes"

[Gman0909]

I would like @colliand 's input in this discussion, as the tags will be used in community-facing initiative bundles that need to be useful as sales assets.

Meanwhile I have revised my original set of tags to follow a verb-noun format:

analytics > expand-analytics
cost-mitigation > reduce-costs
knowledge-sharing > improve-knowledge-sharing
vendor-agnosticism > expand-infrastructure-choice
3rd-party-integration > integrate-3rd-party-solutions
resource-management > extend-resource-management
access-control > refine-access-control
platform-resilience > strengthen-infrastructure-resilience

[aprilmj]

I'm doing some focused work on consortium/coalition/crew design in #7801 and will look at these themes as ways to group communities with common problems. Feedback coming soon!

[aprilmj]

I went back to some very basic personas to think about how we might gather a first consortium, and what we've already learned as part of our problem sensing work.

Here's what I came up with as "themes" (defining a theme as a common problem shared by a consortium, where the theme is made up of multiple initiatives on our roadmap):
Reduce Cost - controls & analytics
Reduce Cost - infrastructure choice
Groups/Access *I'm least sure about this; I think I'm squishing a couple of different use cases together with this one
Knowledge Sharing
Learner Experience
Foundations / Infrastructure Resilience

I came up with this list by thinking first about what member communities have already told us they struggle with. The phrasing would be better if they were all simple Verb noun. sentences.

Some common priorities for science communities (to be refined):

• the funding situation is chaotic; I want my team and my users to be as smart as possible about resources so we stretch every $ as far as possible (Reduce Cost - controls & analytics)
• see above - is there a cheaper way to do what we do? Let’s try it. (Reduce Cost - infrastructure choice)
• see above - some of my friends have funds and want to hop on my hub. I want to welcome new institutions/groups to participate, which makes it important for me to be able to give different groups different access, resources, etc. - and then manage costs for all of them (Groups/Access and Reduce Cost - controls & analytics)
• my users want to collaborate, share and generally read, reproduce or pair on output (Knowledge Sharing)
• and some subset of my users are doing intense compute work; I want them to have the power they need with (see added jc-test hub #1 yet again) the least expense possible (Reduce Cost - infrastructure choice / emphasis on GPU initiatives)
• at the end of the day, we need infrastructure to just work, and want to be good OSS stewards (Foundations)

And for universities / educational institutions (also to be refined):

• as a faculty member or technology provider, I’m expected to show results from the money I spend on infrastructure - I need data around cost, usage and engagement (Reduce Cost - controls & analytics)
• see above - is there a cheaper way to do what we do? Let’s try it. (Reduce Cost - infrastructure choice)
• my institution gets value when students get value, so I want to remove as much friction as possible (Learner Experience)
• different classes have different needs; I want to manage costs and experiences differently for different groups (Reduce Cost - infrastructure choice / emphasis on GPU initiatives, Groups/Access)
• we also need infrastructure to just work, and want to be good OSS stewards - sometimes we even have students who can contribute without significant guidance (Foundations)

[aprilmj]

Sorry for the wall of text! I didn't want to make you read my document just to understand why I proposed that list. But then I just copied half the doc into a GH comment.

Unsurprisingly I think the first consortium should be around "Reduce Costs", combining both

• Reduce Cost - controls & analytics
• Reduce Cost - infrastructure choice

[Gman0909]

Based on the document, I have gone through and added a theme:improve-learning-experience label to our initiatives repo and applied it where I thought appropriate, namely anywhere where either performance, documentation or UI improvements would suggest an overall betterment of the learner's experience of the platform.