cjyabraham
commented
3 months ago Notes from our meeting on Aug 8:
- we want to stand up a site quickly with some dummy architectures so that we can put it in front of end users and get feedback
- we'll start with the codebase from the Glossary site with a single taxonomy of tags and the filter by tag UI
- we'll install Mermaid to help standardize the diagrams and keep them in code and versioned
- we'll start with just the English language but will leave the door open for other translations later
- once we have a sense of the content we are hosting, we can look at some design flourishes to make the site unique later
- I will get a site set up in the next few days and then others can start adding dummy architectures to get a POC ready for feedback
- it's not clear if we need multiple taxonomies at this point or whether we'll just use one. Some examples of ways of tagging architectures is below:
ML Platform
- platform: kubernetes, swarm, mesos
- ML engine: kueue, kubeflow
Financial
- platform: kubernetes, swarm, mesos
- service mesh: istio, linkerd
- supply chain: distributionv2, intoto
They’re going to be using it to host reference architectures, and they’re looking for a similar approach to what we have for TAG sites. The big difference from what I'd say a regular TAG site would be is theres going to be a lot of extra metadata. Essentially a user will submit a page describing their reference architecture, what projects etc they've used & their business, like are they in finance and have 10k+ employees. These should be filterable with some predefined categories (domains).
So we want the accessibility of a static site where people can write up long docs in markdown and manage content via GitHub. We want to support several expanding taxonomies. And we want a powerful filtering UI similar to https://www.cncf.io/case-studies/