HNHacker News
TopNewBestAskShowJobs

glkindlmann

105 karma · joined May 28, 2025

submissionscomments
glkindlmann··on Tetrachromatic Vision
(in the awesome paper shared by varunneal, the metamers are named "keef" and "litz")
glkindlmann··on Tetrachromatic Vision
This is so cool. For your figures, how did you decide the RGB colors of the 4D colorspace? Or did you convince ACM to print your paper with special inks? :)
glkindlmann··on Tetrachromatic Vision
afaik not based on standard RGB displays. All widespread technology for digital color reproduction is based on RGB primaries, i.e. a 3D space of color, or rather a 3D submanifold of spectra inside the effectively infinite-dimensional space of spectra. It is feasible to test for color deficient vision (deficiency or absence of one or more cones, reducing color perception to a 2D or 1D space) because it is easy to sample 3D RGB space and behaviorally detect if colors that are different in 3D are conflated because in some viewer they project to the same location in their 2D or 1D "color" sub-submanifold.

But we'd need a convenient way to sample a 4D space of colors (perhaps with 4 monochromatic sources?), and thereby generate different spectra that normal trichromats see as the same color (called "metamers"), but that tetrachromats could recognize as distinct. And, how the 4D space is sampled would have to be pretty carefully optimized to generate distinct spectra that have the same response with the M (medium or "green") and L (long or "red") cones (which are actually quite similar already!) while also generating different responses for the putative tetrachromat's additional code between M and L. And that isn't possible with any conventional display device.

glkindlmann··on Mathematical Illustrations: A Manual of Geometry and PostScript
Like others[1] I was saddened that Mac's Preview program stopped opening .ps/.eps files with the release of Ventura in late 2022. I never understood exactly why; I guess security issues around the PostScript interpreter. But that made it lot less convenient to hack around in .ps files. Ghostscript's interactive display program(s?) never seemed as convenient as Preview.

Still, this book is awesome and I've been inspired by it for awhile. I used to love whipping up diagrams and simple vector art in .ps/.eps files. In a fit of procrastination in ~2006 I created a minimalist horizontal rule for an IEEE conference submission, and that diamondrule.eps[2] has now become the standard. I would say that EPS is still more convenient than SVG as a format to contain, in one self-contained file, raster graphics with overlaid vector graphics.

[1] https://tidbits.com/2022/10/27/preview-in-ventura-drops-supp...

[2] https://github.com/ieeevgtc/tvcg-journal-latex/blob/main/dia...

glkindlmann··on Riding high in Germany on the world's oldest suspended railway
I learned about the Schwebebahn from this[1] YT video (from The Tim Traveller), which goes into the history of why it made sense to build a suspended rail at this particular place and time. A more recent video [2] gives a much more detailed history.

[1] https://www.youtube.com/watch?v=9IFh6wFTJiQ

[2] https://www.youtube.com/watch?v=sI5DehAuT2I

glkindlmann··on Apple introduces a universal design across platforms
It does indeed feel like a step backward - I was also weirdly reminded of the Forstall skeuomorphism era of UIs.

The video says: "It beautifully refracts light, and dynamically reacts to your movement, with specular highlights"; ugh, why? Why add dynamic==distracting high-frequency details that supply zero information?

The recent super flat UI aesthetic bugged me for awhile for its apparent lack of affordances, but when used consistently it made sense. Now it seems we still get zero affordances, but also visual noise.

glkindlmann··on Show HN: Fast Random Library for C++17
Lengthy indeed. Do you have advice for how to find amongst these the first paper on correctly sampling [0,1) with real-world floating point?
glkindlmann··on Show HN: Fast Random Library for C++17
Now that I've actually looked at the "utl::random" code in the OP, I see that its UniformRealDistribution is a wrapper around std::generate_canonical, so the juicy bits about turning a random into into a random float are not exposed here at all. But the utl::random code does include an pointer* to an informative C++ working group note.

* https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2023/p09...

glkindlmann··on Show HN: Fast Random Library for C++17
Thanks for the additional pointers!

I like the idea of looking at the histogram of exponents: incrementing (by 1) the exponent doubles the width of the interval so should double the number of hits. Conversely the histogram of the significand (or any subset of its bits) should be flat.

glkindlmann··on Show HN: Fast Random Library for C++17
That page ends with "This is the first efficient algorithm for generating random uniform floating-point numbers that can access the entire range of floating point outputs with correct probabilities". I am very skeptical.

The same basic idea is described here [1] (though without any pretty pictures, but with a link to working code), and I know others have implemented the same insight [2].

This seems like the kind of thing that is reinvented every time some who cares about random numbers learns how IEEE 754 works.

[1] http://mumble.net/~campbell/2014/04/28/uniform-random-float

[2] https://github.com/camel-cdr/cauldron/blob/a673553dd7925b0f1...

glkindlmann··on The wake effect: As wind farms expand, some can ‘steal’ each others’ wind
The word "rights" surprisingly didn't come in that piece. By analogy to "water rights" [1], "wind rights" are a thing, both in the basic sense of permission to extract wind power from some chunk of land [2] and the messier sense of that article: conflicts between upstream and downstream users of the wind [3] (recent article and fascinating read)

[1] https://en.wikipedia.org/wiki/Riparian_water_rights

[2] https://en.wikipedia.org/wiki/Wind_rights

[3] https://eelp.law.harvard.edu/wind-wakes-and-the-right-to-win...

glkindlmann··on Beware of Fast-Math
That sounds neat. What would be really neat is if the language helped to expose the consequences of the ensuing rounding error by automating things that are otherwise clumsy for programmers to do manually, like running twice with opposite rounding directions, or running many many times with internally randomized directions (two of the options in Sec 4 of *). That is, it would be cool if Rust enabled people learn about the subtleties of floating point, instead of hiding them away.

* https://people.eecs.berkeley.edu/~wkahan/Mindless.pdf

glkindlmann··on Beware of Fast-Math
Forgive them. That kind of educational failure is common: floating point is just not covered in basic intro classes, or covered hurriedly. There are enough subtleties to how floating point code works that even when an elective class covers it, it is still peripheral to the focus of the class (e.g. intro computer graphics). So students can get by without really coming to terms with the ubiquity of rounding error, or the relationship between denormalized and normalized values. Understanding those fundamentals is a pre-req to understanding how corners are being cut with fast-math.
glkindlmann··on Why are 2025/05/28 and 2025-05-28 different days in JavaScript?
This was an informative (for me) explainer for what's happening in that video https://medium.com/dailyjs/the-why-behind-the-wat-an-explana...
glkindlmann··on Running GPT-2 in WebGL: Rediscovering the Lost Art of GPU Shader Programming
a decade does feel like an eternity here; I remember this book* when it came out

* https://www.oreilly.com/library/view/gpgpu-programming-for/9...

← PreviousPage 2 of 2