Inconsistent ordering of anatomy slot and parts

[lukewarlow]

What's the correct way to make anatomies. Should slots be inside of parts or vice versa? Or is it dependent on context?

I ask because currently the various anatomies aren't consistent in this regard

[gregwhitworth]

Posted this on discord, and will replicate it here:

This somewhat depends on issue 702. However, whether it ends up using slots/parts is actually not as important as doing the research to understand the most common usecases and the necessary parts to unlock those.

Here is a breakdown I did for a potential update to input type=range which when paired with usecases you can arrive at a general recommended anatomy. https://open-ui.org/components/slider.research.parts/

With regards to your specific question however, it depends as the slot is meant to provide flexibility. So you may have nested slots due to being within another part whereas there's another scenario where that isn't necessary.

[lukewarlow]

Adding this to the agenda to discuss. There's examples throughout the site that seem backwards to me

The below being an example of the file structure currently and what I think it should be. This better matches "select-anatomy.jsx" too

Current:

Proposed:

[css-meeting-bot]

The Open UI Community Group just discussed Inconsistent ordering of anatomy slot and parts, and agreed to the following:

• RESOLVED: Slots should wrap parts where the intention is to replace that part as a whole

The full IRC log of that discussion

<keithamus> scribenic: keithamus
<keithamus> scribenick: keithamus
<masonf> q+
<gregwhitworth> q+
<keithamus> Luke_W: throughout openui we've got various anatomies. Parts, slots. We're moving away from the slot idea but it makes sense to document as such. There's inconsistency though - file select has part, then slot inside. Intent is slot, default (which is the part), then you can replace by putting slotted content in. So it's the wrong way around?
<keithamus> masonf: proposed looks better.
<keithamus> gregwhitworth: +1 too. Technically they're slots but it's loose. Placeholder space to insert content. Could be both.
<jarhar> q+
<gregwhitworth> ack masonf
<gregwhitworth> ack gregwhitworth
<keithamus> Luke_W: Replacing text content... but if you replace the whole part...
<keithamus> masonf: slot is on inside... you'd have a button inside a button, which makes no sense. We do have cases of nested slots but that's different
<gregwhitworth> ack jarhar
<keithamus> jarhar: clarify; we're moving away from using the slot attribute but they're still slottables, imperatively assigned
<Luke_W> Proposed Resolution: Slots should wrap parts where the intention is to replace that part as a whole
<masonf> +1
<Luke_W> RESOLVED: Slots should wrap parts where the intention is to replace that part as a whole