Andy Allan

> Meanwhile maybe @gravitystorm will let us know what he had in mind. It was just a general point. Moving complex logic to models makes it easier to test, among...

I should also say that I don't like the scope names suggested in this PR, since I'm concerned that they might be misleading. For example: * `trackable_ordered` - this scope...

> I was wondering about usernames, but they are displayed in a lot of places, and often in contexts where I'd be quite worried about the layout effects of them...

Some edge cases that I've spotted that could be resolved: * We use `zh-hk` whereas OCI uses `zh-HK`. We need to either map between these on import, or fix one...

> The docker run is failing because chromium has in fact been moved to snap only in 22.04 but snaps can't be installed in a docker container so that's that...

> I'm wondering why this change didn't go into small.scss, if it's mainly targeting small screens/mobile devices. I'd like to get rid of small.scss entirely at some point. The way...

I'm not going to argue the merits of one grid system against another - I'm not an expert in either! I'll note that bootstrap 5 has experimental support for CSS...

> You _don't care which rows and columns they occupy_. Why would you go and write out those rows and columns if you don't care about them? We've used https://getbootstrap.com/docs/5.1/layout/grid/#row-columns...

> I still don't understand if you are against specific widths _in pixels_ or widths in any units. In general I'm against specific widths appearing in custom.css (in fact, I'd...

I'm not super keen on this, since browsers are in charge of dealing with language requests. There should (in theory, I know I know I know) be no need for...