From the perspective of someone who has seen a shit ton of the admin notices mess in WordPress, it seems to me the Theme Review Team might be ahead of its time here. The UX issues it tries to tackle by enforcing core-style admin notices partly go beyond its responsibilities; trying to solve them on the theme level, in my humble opinion, could end up doing more harm than good for users.
The one example for an obtrusive notice provided in the team’s meeting was the onboarding notice of the Storefront theme, and the Tavern article provided two more examples of the same type of notice.
Trying to force the implementation of onboarding dialogues for more sophisticated themes into core-style admin notices seems like trying to put out a dumpster fire by, pardon my language, peeing on it – an initiative based on good intentions but incredibly short-sighted.
There is a clear responsibility of WordPress core to provide a more granular API for notices – one that enables visual prioritisation of system-relevant notices over advertising/marketing/upselling. The case for moving notices from their current slot in wp-admin to a dedicated Dashboard Widget and Toolbar item (thus making them less intrusive and more user-centric by default) seems pretty self-explanatory as well.
At the same time, the way of how the WordPress product ecosystem has implemented plugin and theme user onboarding via notices up to today sounds like a deafening cry for a dedicated onboarding API in core.
WordPress has benefited from the incredible diversity in its ecosystem but if that diversity still translates into everybody rolling their own UI within wp-admin in 2019, we’re offloading a key design responsibility of WordPress core to the people who rely on it – both, users and product providers.
I believe this is the underlying concern of the TRT, and I’m all for unifying UX in wp-admin, but as far as I’ve read their meeting notes, the TRT hasn’t provided any compelling argument for effectively banning user onboarding via admin notices aside from the arbitrary notion that bigger notices are undesirable somehow.
I’d like to counter-argue that limitations on how to use an existing notices API won’t solve the problem of a missing API for product onboarding.