robryk @robryk@qoto.org

@enablelanguages I think that people would not consider e.g. the current German flag and Nazi German flag to be different versions of the same flag.

@_thegeoff I'm really curious how uniform will the thickness of the glass be. (I know glass bottles are made that way, but I always assumed that works by careful preforming of the hunk of molten glass that's blown up in the form.)

@_thegeoff Ah, and it's obviously convex (because otherwise how would you extract the final product).

@_thegeoff How are you going to remove the print from within afterwards? Also am curious why would you want the mould to be out of concrete -- it's neither flexible (like silicone) or easy to form (like casting sand).

@grimalkina

Ah, so you're thinking of it as a warning sign, in that unexplained drops (or maybe unexpected changes) suggest an undiscovered side effect of an intervention?

@grimalkina Why would we want to increase employee engagement? I don't doubt that it's correlated with things we[1] actually want, but ISTM that it's a poor proxy for them.

[1] regardless of whether "we" is "the company owners", "employees as a group" or any similar entity

@mjg59

I wonder if adding a flag to openat which prohibits dotdot in the path argument itself would be helpful. It's probably way harder to plumb and obviously provides no way to deal with preexisting code, but provides a way to be careful that's much cheaper in terms of amount of coding/thinking/additional dependencies than what's available now.

@mjg59 I think it would be a bad idea in long term.

It's pretty common to join together pieces of paths from different sources. This case and the symlink case are more-or-less-legitimate situations when a program that would wish to set that prctl might need to deal with dotdot.

Given that we're talking about per-process setting, libraries would need to be able to deal with living in a process that has the prctl enabled. Many libraries don't have knowledge about provenance of paths they manipulate, so the most expedient way of ensuring they work is to do dotdot resolution themselves.

This makes bugs where dotdot resolution can be controlled by the user harder to detect, as the path is not passed to a syscall. It also causes libraries to wish to perform symlink resolution "manually", which works around protected_symlinks sysctl.

@_thegeoff

Yeah, you can see how many assumptions one must make to have students model convection: https://ipho.olimpicos.net/pdf/IPhO_2008_Q3.pdf (for example, the convection cell size is clearly mostly a property of the fluid, and yet I've never found any intuitive explanation for what the size should be or even any Reynolds-number-like symmetries).

Hmm... somewhat relatedly: can one get a conservation law from the symmetry that gives the Reynolds number (via Noether theorem on presumably some weird space that includes fluid properties)?

@_thegeoff

But they would impart only angular momentum to the pot, and not momentum (because they leave the surface level roughly constant and unchanged if averaged over horizontal circles a cm or two in radius). So, this explanation could be tested by letting the pot slide ~frictionlessly sideways and seeing whether its center of mass moves sideways, right?

@_thegeoff

Ah, and I think the rocking would start before major bubble-implosion-noise and clearly visible bubbles (and I messed up the explanation I was thinking of at that time -- the density changes would come from thermal expansion as opposed to steam creation).

@_thegeoff

But why would the liquid start roiling in the first place? Liquid surface tilt and pot tilt are obviously coupled resonators, so practically you will get both oscillating if you get either.

@_thegeoff

I don't think it rocked less, but we didn't really test it for too long, because unless you held it it would splash its contents out at a surprisingly high throughput.

I really should find a pot to break. :)

@_thegeoff

That is a reasonable alternative hypothesis (that we basically get white-ish noise that pushes a dampened pendulum around, so the pendulum will end up with some amplitude of resonant oscillations). It should be easily verifiable if I had such a pot (basically insert a membraneless speaker with the axis oriented radially, play white noise, and see what happens). Perhaps I should find an old pot and reshape it that way with some wood and a hammer.

I realized that I'm actually not sure how a harmonic oscillator driven by noise from e.g. a Wiener process (but without damping) should behave (would its amplitude diverge? probably not).

@_thegeoff who specifically might find it interesting

Consider a pot with a convex bottom (i.e. it will stand on its midpoint) with sufficiently large radius of curvature to make it stable when filled to any level. When you put such a pot on an electric hotplate and get it boiling, it starts rocking: https://www.youtube.com/watch?v=vB_szsLa3zk

I have a hypothesis on what's going on (the pot rocks away from the area where the boiling is more intense due to density difference, contact with hotplate increases rate of heating, so there's more heating on the side that's currently lower and thus any rocking gets amplified). Sadly, I don't have the pot anymore and didn't thing at that time of any experiments that could falsify this hypothesis.

@_thegeoff

How does that work? Is that a polymer/something with different mechanism of adhesion to concrete vs internal cohesion?

@grimalkina

A particular example of same-mold thinking I was thinking of was wanting a metric for "how well a given problem is solved right now" and expecting it to be slowly varying over time, which ignores problems where any partial solutions are worthless or nearly worthless.

@grimalkina

TBF people often try to work under the assumption that the problems businesses solve can be squashed into the same mold too. In my anecdotal experience assuming this and assuming that everyone works the same way is correlated, so much so I assumed without thinking that they have the same cause.