The AI crank always cracks me up.
Of course, glad to see it was another @isohedral project.
But then, when trapped in a local maxima prohibiting growth, pressure builds as too many new layers attempt to shim themselves under existing layers, until inevitably the stack collapses somewhere.
Then new layers can restart generating new apex baby layers on a now higher foundation of fertile fragmented but compressed and stable new-legacy rubble. Another point-oh age begins.
And sometimes, the stack just falls apart because.
In between those extinction events, layers that spawn the most layers, and form opportunistic bridges over lateral layers, dominate and thrive.
Occasionally, some layers try to reorder themselves to optimize future growth. Or tunnel down to achieve stronger footing. But like the tower of Hanoi, the more layers involved, the more intractable the replanting and reordering. Meanwhile, other growth routes around them. Yet, many instances of these failed structures can be found in the depths.
The world keeps moving around us. Can't choose staying still.
https://suvakov.github.io/vibes/SlidingPuzzleChess/index.htm...
Lovely idea by the way.
"The Internet Was Weeks Away From Disaster and No One Knew"
https://www.youtube.com/watch?v=aoag03mSuXQ
(at about the 9:50 mark)
The gravitational constant is maybe a little low for my taste, but I like that I can fling a block vertically up off the top of the frame and it reappears even 5+ seconds later. Things don't get ignored out of existence. Neat.
I know the time it takes to get something to feel this good.
Really fantastic work.
You'd first need to figure out a way to generate a complete dependency tree. For each box, I interpret its height as a measure of its complexity and its width as a measure of the support it receives. The hardest part would probably be figuring out a way to quantitatively measure those values.
Previously I'd postpone some tooling since I'd lost more time on it (unless it's something I wanted to learn anyway), but now I'm all in.
When did things get specialized this much?
Thank you for the laughs. I needed that!
for (let re of rects) {
push();
translate(re.body.position.x, re.body.position.y);
rotate(re.body.angle);
- rect(0, 0, re.w, re.h, 2);
+ rect(0, 0, re.w - 1, re.h - 1, 2);
pop();
}Things are much nicer now and the problem is entirely avoided by using pointer events: https://developer.mozilla.org/en-US/docs/Web/API/Element/set...
Something similar happens all the time in games when you go from a static version of something to the higher level of detail version with physics enabled, if the transition isn't handled gracefully or early enough you can get snapping.
It's the single pin under everything because there are a limited number of those cables especially in some regions so a single shark can take out the entire internet for some countries.
I think that's the other metaphor here.
It's not just standing on the tiny shoulders of one forgotten maintainer. The entire system only appears stable because we're looking at a snapshot of it.
In reality it's already collapsing.
I've been using it so long, I forgot that it is not official.
I don't know if the "The Nebraska Guy Ranking" this project uses is very useful, though. In particular the "depth" criteria doesn't make much sense to me, since it assumes the more foundational a dependency is, the more robust it must be. This seems to run counter to the point of the original comic where the "Nebraska Guy" piece was the fragile block holding up the entire tower.
This project also doesn't attempt to measure or visualize the complexity of a project. Theoretically a more complex project would require more support than a simple one, so I think that's an important metric to capture.
Edit oh and Extrapolating out; 605.
There are many regions that are served by a single line, more than you think.
Even "well connected" places have fewer cables than you expect, and the frustrating thing is that you don't know that you can route around an issue until you try.
BGP is really resilient, which is great, but if your path is not clear then you'll only realise it when the failover doesn't happen, you'll think there's a redundant path.
For small island countries and such, satellite capacity may be sufficient; and it is likely helpful for keeping international calling alive even if it's not sufficient for international data. But when you drop capacity by a factor of 1000, it's going to be super messy.
This is never going to change because from a physical perspective free radio is a shared medium while each individual fiber (or wire) has its own private bandwidth.
I’m changing this ASAP to least-privilege and I’ll publish a clear explanation of scopes + data handling. In the meantime: please run the local/CLI path if you want zero-trust.
[0] https://www.google.com/maps/d/viewer?mid=1805q6rlePY4WZd8QMO...