Profile avatar
ledcoyote.bsky.social
engineering faerie bug https://ledcoyote.com
53 posts 43 followers 65 following
Regular Contributor
Active Commenter
comment in response to post
old fashioned cake donut hell yeah
comment in response to post
Reminds me of the Bluey episode where Bandit takes the kids to the pool by himself and forgets all the basic shit and Chilli comes along and saves the day youtu.be/YpI0jgqNJGc?...
comment in response to post
Okay, yeah, greater or less than 50 seems to be the goal, thus I would have recommended thresholds. The weird gradient threw me off, and I'm sad I spent more brain cycles on it than I should have. It would be cool to not abuse color like this 🫶
comment in response to post
Or if <50 vs >50 is the goal here, a binary threshold
comment in response to post
Would not a uniformly varying gradient serve better?
comment in response to post
I have questions about your color scheme. It *very strongly* accentuates differences in the 47-50% range while flattening differences above 50%. What is the purpose of this choice? It weirdly calls out WY and ND, though they're barely different from WV. And 51 to 70% is barely noticeable 🤷
comment in response to post
Of course that argument makes them look bad, so instead we get it by innuendo: Lee killing US soldiers to support his state's defense of *slavery* is 'good,' but folks in LA protesting in support of their state's defense of immigrants is bad. Left unstated: because racism.
comment in response to post
Another unexpected result of the mysterious and arcane math involved here, when you could previously spend $100 to get 100% of a thing, you will now spend $100 to get $100÷$130=76.9% of a thing, meaning your money goes 23% less far!
comment in response to post
I'll go out on a limb and conjecture: 3.3uW is the power draw for all non-maths "support" functionality, active alone most of the time, thus the "typical" draw, while 10.5uW-3.3uW = 7.2uW *is* the power required for the math itself—clocking, register juggling, adders, etc. Thus 7.2uW*0.15s = 1.1uJ
comment in response to post
Is a good rock
comment in response to post
why would you manually kern caps Times New Roman based on the serif edges??? I’m gonna lose my mind and so is god
comment in response to post
Hmm, 3.3uW appears to be the quiescent (standby power) given the supply current in WAIT mode 2.2uA x 1.5V. In OPERATE mode, the draw is higher, 7uA x 1.5V = 10.5uW. So my crude calculation above becomes 10.5uW x 0.15s = 1.6uJ 🤷
comment in response to post
And 3.3 microwatts includes *all* the functions of the IC, including I/O like capturing button presses and segment display, managing memory, and so on. The amount used for only the maths will be less, probably a lot less.
comment in response to post
Page 7 shows the duration of various operations, a basic multiply taking 100~200ms. Crudely, 3.3e-6 x 150e-3 gives you 0.50 microjoules. Getting more precise will require the I/V waveforms clocking through the registers over this calculation, which we don't have here. But this is pretty close.
comment in response to post
Apropos: bsky.app/profile/karl...
comment in response to post
Lol, I came here to say this
comment in response to post
And thanks for all the fish!
comment in response to post
Aw hell yeah. I love it when people notice and recognize their awesomeness ❤️
comment in response to post
Appending /g will replace all instances instead of just the first 😃
comment in response to post
I thought of @justingarrison.com and @withenoughcoffee.com as I was writing this. I just found your new podcast and have to catch up!
comment in response to post
Cool! Thanks for this history. I really liked this film