I previously managed a firewall via scripts which would automatically revert your update in 20 seconds unless interrupted. So if you botched it and lost access, you just had to sit tight for 20 seconds.

On Reddit perhaps. Not here.

gweights

14% over estimates it because the user isn't clicking with uniform randomness, their clicks are normally distributed about the center of the line.

>In total the thickness went down from 7 to 6 pixels, which is a 14% decrease, making it 14% more likely to miss it.

Pedantic, but chance of miss is actually less than 14% more likely since the user's click location is not uniformly random over the thickness area, it's biased toward the center (normally distributed).

Pedantic, you don't know the distribution, so the chance could be higher

The reduction was specifically to the in-window side of the edge, so it's definitely greater than 14%.

Interesting, I've always approached from the outside in.

I approach from whatever side the mouse happens to be on...

never thought about it before but after playing with it a while i notice i tend to approach from the right, which means moving out if i'm inside on the right side. i think this is because my positioning accuracy seems to be higher moving leftwards than rightwards...

We can safely assume they're more likely to be close to the edge they're trying to grab than some random location on the window

Aim wider: why window and not screen?

Yeah, and not to mention the increase in likelihood click events the user intends for the application will make it through successfully, rather than being stolen by the window manager.

Technically correct is the best kind of correct.

I had similar thought but didn't want to be that guy.

My take is sometimes we get paid to be that guy and precision has its place and value.

We get lost when being right is seen as having value - instead of improving clarity and precision if needed in a specific context.

No furniture in front of the window.

See https://Schematix.com

Has a graph query capability which is like a reasoning engine. And beyond that, can run simulations on the models.

Multi user, version control with branch and merge, time shifting, etc.

It's for diagramming whole environments, not just discrete diagrams.

This is incredibly cool! I've been using Neo4j to hack something like this together but the UI on this is amazing. I honestly believe that tools like this should be as fundamental to IT as double entry bookkeeping is to accounting and CAD is to engineering and blueprints are to construction.

How many died waiting for approval of cures we did find?

How many have died because we didn’t care enough to try harder to cure these terrible diseases. Tens of millions!

If you have a disease and want to find clinical trials for unapproved drugs, they are available.

https://clinicaltrials.gov/

https://clinicaltrials.gov/study/NCT05855200?locStr=New%20Yo...

Far less than if we didn't have a proper system in place. We literally have miracle cures for things like measles, polio, and small pox yet people are far too stupid to take them. You think just letting people try whatever they want on people in a desperate situation is going to lead to good things?

If you like diagrams that are also interactive via topological graph queries, See:

https://schematix.com/video/depmap/

Kinda cool looking product, I like its perspective a lot especially the 'impact' view you can do. The physical view is something unique as well.