Closed adambraimbridge closed 5 years ago
Biz-ops is meant to be mostly un-opinionated about the data it stores.
On the other hand, Tako (next-takocat) is meant to be opinionated about how it knows which repositories are Next, how they're grouped (into, e.g, apps / components / other), and so on.
It's unreasonable to code all of that business logic into the biz-ops system.
It's possible that in future next-takocat can push information to biz-ops whenever its repositories change.
Who says that biz-ops is meant to be un-opinionated? The data layer is flexible so you should be able to shove as many opinions into it as you want. We don't want to put next-specific business logic into the biz-ops view layer, but there's nothing stopping you from using its data layer as your canonical data source and annotating it with your own custom details.
We don't want to put next-specific business logic into the biz-ops view layer, but there's nothing stopping you from using its data layer as your canonical data source and annotating it with your own custom details.
Okie dokie, the question then is how do we do that? Specific requirement would be to surface the topics of a repository via the API.
I think our rough plan is going to be to continue with this, because then we can agree on a shared schema for our tools, but where this gets its data from can/should change in future.
/cc @lucas42 @wheresrhys
Here's a rough stab at getting topics: https://github.com/Financial-Times/github-importer/pull/10
We had a IRL meeting about this, see https://financialtimes.slack.com/archives/CD689QU2K/p1541499781012800.
We can define ownership of next repositories as the list of repos that
next-tako
managesWe could also define ownership of next repositories as the list of next repositories from Biz Ops
~We should explain why 1 is better than 2.~