Ask HN: In your experience, what are sound conventions for e-ink UI development?

I responded to one of your surveys: dive in to supporting "xbox + index.html + trmnl.js" as a BYOD / demo target! s/xbox/raspbpi/...etc... I've got extra monitors and RPI's dangling around, and it's not out of the question to splash $50 for a BYOD license and get a "free" TRMNL display that I could have running during daylight hours, or wired to a smart outlet or something?

Also: wouldn't it be great to have an "advanced" timeline customizer, eg: lambda-ish functions? (this I didn't bother sending to y'all...)

I've generally got mine wired to `Calender(15m) => Weather(5m) => Calendar(15m) => MotivationalQuote(5m) => Calendar(15m) => ...etc...`, but that's extremely gross (as you can imagine) with a bunch of `Calendar (copy) (copy) ...` because I can't be bothered to spend time fixing it up.

Thinking it through, I'd kindof like "primary queue (15m)" vs "secondary queue(5m)" and have "the algorithm" automatically merge them at runtime (ie: two pointer heads as you alternate queues), but the end game is effectively:

    def next_card(handle):
        if handle.time.hour == "12:00":
            return ( "Lunch Break", "30m" )
        elif handle.time % 60 == 0:
            return ( "Calendar", "15m" )
        else:
            return ( handle.plugins.shuffle()[0], "5m" )
        ...etc...

<3 ... the AI stuff looks tasteful + cool, please thank your support staff, and excited to see where y'all end up going! (Reminds me, need to figure out the supposed firmware update...)

I love this question. Most eink UIs feel like they just took interaction patterns from smartphones and crammed them in. I would love to see more innovation on eink UI patterns that works from first principles given the nature and limitations of eink!

For example, I consider dragging operations an anti-pattern on eink devices. Both latency and touch precision are just not at the levels we expect from years of using phones, leading to an experience that is prone to errors and frustration. A particularly common problem is messing up your state just as you lift your finger. I can't tell you the number of times I have carefully highlighted a passage only to have it jump to the wrong words as I release. Dragging is also just an inherently physically awkward maneuver, especially when performed prone.

I think a better pattern would be to use discrete mode changes with modal controls for refining state. As a simple example, highlighting could work like this: 1. Touch and briefly hold to enter selection mode. 2. Selection mode brings up large buttons around the current selection to refine it. Refinement happens via taps, not drags. 3. Selection mode is exited via another button tap. Confirmation is explicit, not implicit. This feels like it would be a lot more relaxing and less error prone.

I could also see a hybridized approach for interactions (like highlighting) where refining exclusively via button taps might be tedious. In this case you could use drag gestures while in selection mode to quickly add/remove large sections to/from the current selection state. This would give you a shortcut for highlighting large sections quickly, which could then be detail-refined via button taps. Having it be modal would also allow you to undo any accidents.

A further refinement for this interaction pattern would be to offer intelligent controls that give the user context-inferred options for their refinement. For example, if you have part of a paragraph highlighted, a button could appear to highlight the rest of it, and subsequently buttons could appear to highlight the adjacent paragraphs. This could even be combined with the hybrid drag gestures idea - after quickly dragging over a rough area, the system could offer multiple proposed selections if there is ambiguity what the user may have intended.

I have this exact image (plus one of Dürer's hare images) in my 'sleep screen' folder for my larger readers. I also love German Expressionist woodcuts on ePaper, particularly on smaller screens. Drama is high (maybe too dark/bleak for some) and the stark, high-contrast compositions carry over to the tech beautifully.