1. Cassidy's Notes
1.1. High Level Discussion
There are sort of two mental paths: one unified experience that is the same everywhere, or a set of building blocks that downstreams customize/change/build from.
1.1.1. Concerns
- Losing the identity of the Shell as it improves/changes
- Downstreams (Endless, Pop!) are going to have different targets, principles
1.1.2. Things that have worked well
- Search-based
- Distraction-free
- Notifications (to an extent—mobile has surpassed this)
- System Status
1.1.3. Things to Discuss
- Activities Overview
- App picker/launching
- Folders/drag-and-drop
- Window picker
- Window tiling
- Window decorations/titlebars/theming
- First-run experience + discoverability + overlay help + empty state
- Actions in search (like HUD/Plotinus/menus)
1.2. Activities Overview
- Question around whether app- or window-centric is how people think/work
- Or if we can better cater to both
- Effectiveness of big grid of alphabetic apps, customization
- Crafting your own space for work
- Alphabetical may not be helpful
- Folders are non-manipulable
- Full list of apps is overwhelming, shouldn’t be the default/main case
- Pagination is weird, non*spatial
- Discoverability, explorability, first-run
- Search
1.3. First-Run/Overlay
1.3.1. Personal Thoughts
Tutorials are problematic to me philisophically becuase
- User Interface should be self-evident; generally if you require a tutorial, it’s not obvious enough
- It’s a lot of information at once the user is expected to
- See/read in the first place
- Understand
- Remember
- On-demand hand-holding (tooltips or smart, contextual information) is more sticky to the user
- Annoying for experienced users
- Typically only on first-run
- Doesn’t scale well to a multi-user system
That said, I don't think these preclude having an initial first-run overlay, I just don't know it should be the only thing we rely on.
1.3.2. Intel Initial Setup Study
Intel ran a study in a UX lab (like someone's living room)
- People felt let down, wanted a bit more
- Want to interact with it/learn more right away, have something to do
1.3.3. Implementation Possibilities
Not mutually incompatible
- Help Guide
- First-run overlay
- “Feature gallery” in Initial Setup
- On-demand videos/tutorials (Dbus API or something)
- Progerssive disclosure (tooltips)
- Making Help more discoverable (system status area?)
1.4. Lock/Greeter
1.4.1. Original goals
- Privacy
- physical shield protecting your data
- Touch
- place to interact
- way to block stray input
- Useful
- music playback
- notifications
- Personalization (one place you see wallpaper quite a lot)
1.4.2. Things Pop!_OS Changes
- Disable shield on log-in screen
- Turn off notifications by default
- Because displays turn on/light up constantly when people are out of the office, especially when it's a desktop with external display
- Many displays are moving to not having explicit on/off buttons (recent Apple, Dell, LG displays)
- Privacy-wise, it should probably be opt-in instead of opt-out
1.4.3. Things We Like
- Date/time/status
- Music info + controls
1.4.4. Personal thoughts
- Remember last logged-in user
- Don't require a click to "select" the first user (login case should be more similar to lock case)
1.4.5. Friday Morning
- Notifications (or at least the fact that you have them) should be visible while typing your password (if they're before, it means to just start typing to unlock and miss your notifications
- Better efficient use of screen space (more responsive design with landscape vs portrait)
- Guest or Demo modes might be treated differently than a user/person
1.4.5.1. Notifications
- Lock screen is a small amount of time; maybe less potential for exploration of notifications
- May be more valuable to just know the apps that sent notifications, or even just that are notifications (also more privacy-conscious)
- Could algorithmically set default avatar so that avatars can be used as a more recognizable identity (combine hash of user name to select a palette color plus first letter or two?)
1.4.6. elementary OS
- One-key (up/down) to select user
- Want to design for the potential of having notifications (GNOME does this by having the lock screen IN the user's session, but could move to have the user switching part live there too by giving standard users the permission to switch to other users)
- Could be able to assume unique avatars if we use Granite avatar API