Corbin Crutchley
@crutchcorn.dev
VP of Software Engineering | @GitHub.com Star | Microsoft MVP | @playfulprogramming.com Partner | @TanStack.com Maintainer | Twitch Streamer (http://twitch.tv/crutchcorn) 💅💅💅
Thanks for coming to my TED talk.
January 2, 2026 at 9:32 AM
Thanks for coming to my TED talk.
This is why (I assume) the OneNote pen selection is so bad BTW; they're not using the Onyx inking APIs so it's just a slow refresh rate as you saw with the Jetpack Ink APIs.
January 2, 2026 at 9:32 AM
This is why (I assume) the OneNote pen selection is so bad BTW; they're not using the Onyx inking APIs so it's just a slow refresh rate as you saw with the Jetpack Ink APIs.
So if you rely on the dry mode to return you a list of points you can serialize to a DB and then restore later it's a bad time. "Add a bunch of points you didn't expect" bad time.
Why the official Boox apps don't seem to suffer from this IDK. Maybe they always render in a different mode or smth??
Why the official Boox apps don't seem to suffer from this IDK. Maybe they always render in a different mode or smth??
January 2, 2026 at 9:32 AM
So if you rely on the dry mode to return you a list of points you can serialize to a DB and then restore later it's a bad time. "Add a bunch of points you didn't expect" bad time.
Why the official Boox apps don't seem to suffer from this IDK. Maybe they always render in a different mode or smth??
Why the official Boox apps don't seem to suffer from this IDK. Maybe they always render in a different mode or smth??
And, from my understanding, the Boox API has two modes:
- Wet inking (when moving the pen)
- Dry inking (when the pen is done moving)
But this is another problem: The "dry" ink doesn't seem to record the points in the same way that the "wet" ink does.
- Wet inking (when moving the pen)
- Dry inking (when the pen is done moving)
But this is another problem: The "dry" ink doesn't seem to record the points in the same way that the "wet" ink does.
January 2, 2026 at 9:30 AM
And, from my understanding, the Boox API has two modes:
- Wet inking (when moving the pen)
- Dry inking (when the pen is done moving)
But this is another problem: The "dry" ink doesn't seem to record the points in the same way that the "wet" ink does.
- Wet inking (when moving the pen)
- Dry inking (when the pen is done moving)
But this is another problem: The "dry" ink doesn't seem to record the points in the same way that the "wet" ink does.
... And by the fact that you can draw over traditionally non-drawable UI elements 🤡
To solve this, they expect you to pass excluded rects of screen positions to avoid drawing over UI elements. Need rounded corners tho? LOL nope! User can draw around the edges and see they're really square
To solve this, they expect you to pass excluded rects of screen positions to avoid drawing over UI elements. Need rounded corners tho? LOL nope! User can draw around the edges and see they're really square
January 2, 2026 at 9:28 AM
... And by the fact that you can draw over traditionally non-drawable UI elements 🤡
To solve this, they expect you to pass excluded rects of screen positions to avoid drawing over UI elements. Need rounded corners tho? LOL nope! User can draw around the edges and see they're really square
To solve this, they expect you to pass excluded rects of screen positions to avoid drawing over UI elements. Need rounded corners tho? LOL nope! User can draw around the edges and see they're really square
You can see this when you overlay their SDK's drawing "View" with a transparent background and add a colorful Jetpack BG there: it adds weird artifacting around your pen strokes.
January 2, 2026 at 9:27 AM
You can see this when you overlay their SDK's drawing "View" with a transparent background and add a colorful Jetpack BG there: it adds weird artifacting around your pen strokes.
"So why is the Boox SDK 'problematic'" I can hear you ask
Well....
It seems that the way the SDK works is that it interfaces with the digitizer and screen in some kind of direct interface way that bypasses Android's graphics APIs by overlaying the drawing from the firmware directly???
Well....
It seems that the way the SDK works is that it interfaces with the digitizer and screen in some kind of direct interface way that bypasses Android's graphics APIs by overlaying the drawing from the firmware directly???
January 2, 2026 at 9:26 AM
"So why is the Boox SDK 'problematic'" I can hear you ask
Well....
It seems that the way the SDK works is that it interfaces with the digitizer and screen in some kind of direct interface way that bypasses Android's graphics APIs by overlaying the drawing from the firmware directly???
Well....
It seems that the way the SDK works is that it interfaces with the digitizer and screen in some kind of direct interface way that bypasses Android's graphics APIs by overlaying the drawing from the firmware directly???
OK so to be faiiiir: Boox has a (problematic) SDK that allows you to integrate with what I assume is the same underlying API.
I'll loop back to why it's problematic in the next posts, but see this video for why it doesn't even fix all of OneNote:
I'll loop back to why it's problematic in the next posts, but see this video for why it doesn't even fix all of OneNote:
January 2, 2026 at 9:24 AM
OK so to be faiiiir: Boox has a (problematic) SDK that allows you to integrate with what I assume is the same underlying API.
I'll loop back to why it's problematic in the next posts, but see this video for why it doesn't even fix all of OneNote:
I'll loop back to why it's problematic in the next posts, but see this video for why it doesn't even fix all of OneNote:
Very likely, using EDL:
www.temblast.com/edl.htm
But this comes with a number of problems:
1) `onyx_config` isn't simple ASCII anymore and may be MMKV formatted: github.com/jdkruzr/Boox...
2) You lose firmware updates or anything else that Boox might have
3) Risk of device itself
www.temblast.com/edl.htm
But this comes with a number of problems:
1) `onyx_config` isn't simple ASCII anymore and may be MMKV formatted: github.com/jdkruzr/Boox...
2) You lose firmware updates or anything else that Boox might have
3) Risk of device itself
EDL Utility
www.temblast.com
January 2, 2026 at 9:21 AM
Very likely, using EDL:
www.temblast.com/edl.htm
But this comes with a number of problems:
1) `onyx_config` isn't simple ASCII anymore and may be MMKV formatted: github.com/jdkruzr/Boox...
2) You lose firmware updates or anything else that Boox might have
3) Risk of device itself
www.temblast.com/edl.htm
But this comes with a number of problems:
1) `onyx_config` isn't simple ASCII anymore and may be MMKV formatted: github.com/jdkruzr/Boox...
2) You lose firmware updates or anything else that Boox might have
3) Risk of device itself
....
It's because in Boox's firmware (as of 3.2.2) they hardcode OneNote, WPS, and EverNote to use their inking stuff rather than the built-in stuff
...
Dead end. 😭
It's because in Boox's firmware (as of 3.2.2) they hardcode OneNote, WPS, and EverNote to use their inking stuff rather than the built-in stuff
...
Dead end. 😭
January 2, 2026 at 9:07 AM
....
It's because in Boox's firmware (as of 3.2.2) they hardcode OneNote, WPS, and EverNote to use their inking stuff rather than the built-in stuff
...
Dead end. 😭
It's because in Boox's firmware (as of 3.2.2) they hardcode OneNote, WPS, and EverNote to use their inking stuff rather than the built-in stuff
...
Dead end. 😭
OK sorry about the transparency on that image...
But 4 million logs in 5s (apparently???)...
And still latency 🙈
But 4 million logs in 5s (apparently???)...
And still latency 🙈
January 2, 2026 at 8:02 AM
OK sorry about the transparency on that image...
But 4 million logs in 5s (apparently???)...
And still latency 🙈
But 4 million logs in 5s (apparently???)...
And still latency 🙈
Giving GameActivity a shot to see if the bottleneck is JNI input throughput
January 2, 2026 at 7:55 AM
Giving GameActivity a shot to see if the bottleneck is JNI input throughput
Apparently 0.3ms still isn't fast enough...
And I don't think eInk needs 3000fps so maybe time to try a different approach...
And I don't think eInk needs 3000fps so maybe time to try a different approach...
January 2, 2026 at 7:35 AM
Apparently 0.3ms still isn't fast enough...
And I don't think eInk needs 3000fps so maybe time to try a different approach...
And I don't think eInk needs 3000fps so maybe time to try a different approach...
Mine too fwiw haha. I'm just trying to overachieve for @reikaze.bsky.social and make an app that suits his needs well enough haha
January 2, 2026 at 7:25 AM
Mine too fwiw haha. I'm just trying to overachieve for @reikaze.bsky.social and make an app that suits his needs well enough haha
TBH if I can't figure out how OneNote has such low writing latency, I might return mine 😅
I've been lazer-focused on building this app for my friend and the alternatives are poo
I've been lazer-focused on building this app for my friend and the alternatives are poo
January 2, 2026 at 7:10 AM
TBH if I can't figure out how OneNote has such low writing latency, I might return mine 😅
I've been lazer-focused on building this app for my friend and the alternatives are poo
I've been lazer-focused on building this app for my friend and the alternatives are poo
OK entirely rewrote the app using Skia (a different Canvas mechanism) and it's still latency to hecc
January 2, 2026 at 6:39 AM
OK entirely rewrote the app using Skia (a different Canvas mechanism) and it's still latency to hecc
OK I can't find another Android app that runs this well on Boox except first-party apps and OneNote...
But they can't be using Boox SDKs because they don't support 16KB page files:
github.com/onyx-intl/On...
... Time to decomp OneNote ...
But they can't be using Boox SDKs because they don't support 16KB page files:
github.com/onyx-intl/On...
... Time to decomp OneNote ...
Support 16 KB page sizes · Issue #132 · onyx-intl/OnyxAndroidDemo
Please can you make onyxsdk-pen compatible with 16KB page requirements? Apps using this sdk will get rejected from the Play Store without 16KB support: Starting November 1st, 2025, all new apps and...
github.com
January 2, 2026 at 6:13 AM
OK I can't find another Android app that runs this well on Boox except first-party apps and OneNote...
But they can't be using Boox SDKs because they don't support 16KB page files:
github.com/onyx-intl/On...
... Time to decomp OneNote ...
But they can't be using Boox SDKs because they don't support 16KB page files:
github.com/onyx-intl/On...
... Time to decomp OneNote ...
No.... Notein sucks too...
January 2, 2026 at 5:56 AM
No.... Notein sucks too...
Whoa whoa whoa... OneNote works fine... Is this just a Jetpack Ink thing???
January 2, 2026 at 5:53 AM
Whoa whoa whoa... OneNote works fine... Is this just a Jetpack Ink thing???
Turns out that alternatives like Viwood likely also suffer
www.reddit.com/r/viwoods/co...
AFAIK this is caused by refresh issues on eInk where they need to run C code on display/digitizer firmware instead of the OS (and then overlay to the OS) after some deep diving.
Understandable, but annoying.
www.reddit.com/r/viwoods/co...
AFAIK this is caused by refresh issues on eInk where they need to run C code on display/digitizer firmware instead of the OS (and then overlay to the OS) after some deep diving.
Understandable, but annoying.
From the viwoods community on Reddit
Explore this post and more from the viwoods community
www.reddit.com
January 2, 2026 at 5:36 AM
Turns out that alternatives like Viwood likely also suffer
www.reddit.com/r/viwoods/co...
AFAIK this is caused by refresh issues on eInk where they need to run C code on display/digitizer firmware instead of the OS (and then overlay to the OS) after some deep diving.
Understandable, but annoying.
www.reddit.com/r/viwoods/co...
AFAIK this is caused by refresh issues on eInk where they need to run C code on display/digitizer firmware instead of the OS (and then overlay to the OS) after some deep diving.
Understandable, but annoying.