@mattly @polotek yup, native used to be better at this but no more
Conversation
Notices
-
Embed this notice
Jan Lehnardt :couchdb: (janl@narrativ.es)'s status on Friday, 19-Jul-2024 02:06:44 JST Jan Lehnardt :couchdb: -
Embed this notice
Jan Lehnardt :couchdb: (janl@narrativ.es)'s status on Thursday, 18-Jul-2024 23:41:47 JST Jan Lehnardt :couchdb: hot take: having dozens of UI components go through a placeholder/loading/populated state progression on an individual timeline makes for a terrible user experience.
This includes most modern web and native things. I know how we got here, but the flickerfest needs to be a transitional period and I’m not seeing a way out. 🔥
-
Embed this notice
esmevane, sorry (ironchamber@mastodon.esmevane.com)'s status on Friday, 19-Jul-2024 01:44:41 JST esmevane, sorry @darius @polotek @janl Yeah. I get the motivation, too. Saying "just do this thing" (where in this case it's "take proper care to thoroughly model out your async state and design your component space so it' knows about that state") is kind of bad vs "let's make an official API". But man, the API is tough to grok and even tougher to talk about w/ folks.
-
Embed this notice
Darius Kazemi (darius@friend.camp)'s status on Friday, 19-Jul-2024 01:44:42 JST Darius Kazemi @polotek @janl yeah Suspense is sort of trying to mitigate the thing where every single individual async component has its own loading indicator and pops in. It lets you draw a line around components (technically just children of the Suspense component) and have finer control over when the sub-components pop in, what order they pop in, etc.
this is not me saying it's GOOD btw, haha
-
Embed this notice
Marco Rogers (polotek@social.polotek.net)'s status on Friday, 19-Jul-2024 01:44:43 JST Marco Rogers @janl this is what the Suspense thing in react does right? I don't fully understand it. But it seems like it's meant to wait until all async calls are resolved? Either way I find it very confusing.
-
Embed this notice
Jan Lehnardt :couchdb: (janl@narrativ.es)'s status on Friday, 19-Jul-2024 01:54:14 JST Jan Lehnardt :couchdb: @polotek native apps suffer the same fwiw. Terrible database access can be done anywhere ;D
-
Embed this notice
Marco Rogers (polotek@social.polotek.net)'s status on Friday, 19-Jul-2024 01:54:15 JST Marco Rogers @janl guarantees? On the web? Ha!
-
Embed this notice
Jan Lehnardt :couchdb: (janl@narrativ.es)'s status on Friday, 19-Jul-2024 01:54:16 JST Jan Lehnardt :couchdb: @polotek no idea, but blocking the UI until all async things are done is also not great if those aren’t guaranteed to resolve fast.
-
Embed this notice
Matthew Lyon (mattly@hachyderm.io)'s status on Friday, 19-Jul-2024 02:06:43 JST Matthew Lyon @janl @polotek to a broader extent, I think part of this also comes from screens that need to gather data from multiple sources – whether the app is talking to multiple APIs directly or through some sort of “backend for a frontend” proxy
“what do we do in the UI if one of the services is misbehaving?” is a question that requires a technically-oriented designer to have product-direction-level authority to answer, but those people don’t exist in modern software orgs
-
Embed this notice
esmevane, sorry (ironchamber@mastodon.esmevane.com)'s status on Friday, 19-Jul-2024 02:52:58 JST esmevane, sorry @janl @darius @polotek oh shoot. haha welp. so you did! my bad.
also, re: the flickerfest limbo - 😭
-
Embed this notice
Jan Lehnardt :couchdb: (janl@narrativ.es)'s status on Friday, 19-Jul-2024 02:53:00 JST Jan Lehnardt :couchdb: @ironchamber @darius @polotek as I said, I know how we got here :)
-
Embed this notice
Brian David (bcdavid@hachyderm.io)'s status on Friday, 19-Jul-2024 03:11:54 JST Brian David @janl @mattly @polotek Yes, it's mostly the org chart vomiting onto the screen. And also, I have found that all backend devs and engineering VPs I've talked to about this seriously believe this is a better user experience than waiting longer with a splash screen but getting a more complete final render. A lot of people don't realize that multiple shimmers resolving at different times is usually perceived as slower to the user and introduces tons of jank.
-
Embed this notice
Jan Lehnardt :couchdb: (janl@narrativ.es)'s status on Friday, 19-Jul-2024 03:11:56 JST Jan Lehnardt :couchdb: @mattly @polotek just realising this is conway’s law striking yet again.
-
Embed this notice