back
user profile
raphlinus
13,685karma·1,846submissions·March 7, 2014
about
I do research on fundamental UI technology and 2D graphics, with a focus on Rust and fonts.
@raph@mastodon.online
recent activity (1,846 total)
comment
Yeah, it's no walk in the park, let me tell you that.
comment
You are right, and that is an excellent way to state my goals: I hope to create systems that have as much lasting impact as autotools.
comment
I think you nailed it. I wake up each morning consumed with anxiety, thinking to myself, "how do I become relevant?" Clearly, having shipped advanced multilingual typesetting capabilities to…
comment
I was only tangentially involved in X, but was active in a lot of related projects (some Gtk infrastructure, Ghostscript, the seeds of what became Inkscape), and now work on GPU-accelerated font rende…
comment
I think the tradition of April Fools (and related, Onion-like satire) serves an important epistemological function, even more relevant and important in today's information environment. You cannot…
comment
This will be changing soon. I believe Chrome M113 is scheduled to ship to stable on May 2, and will support WebGPU 1.0. I agree it's a game-changing technology.
comment
The specific thing I asked for is very technically challenging. One of the things I can offer is some tests that fail in, among other things, MoltenVk, as the mapping to Metal barriers is imprecise. I…
comment
The big question I have is whether this can possibly support mandatory Vulkan features that are not available in Metal. The one I care about most is device scoped barriers, which in turn are needed fo…
comment
It is. I've been waiting for the brilliant take, the one that makes me say, "wow, this really gets to the essence of the whole thing." I haven't found it yet. It's not the blu…
comment
I'm curious why they haven't done more forensic typographic analysis of the fonts themselves. There are a number of images of actual Selectric fonts, including [1], and that includes symbols…
comment
Three. You can also do it in Metal, which as of recently has cooperative matrix multiplication in the form of the simd_matrix type (this is similar functionality as "tensor cores" in the Nvi…
comment
Another interesting analysis: https://armoxon.substack.com/p/the-case-for-shunning …
comment
Carry is also a monoid. Consider addition of two binary numbers. If the two digits are zero, then the output carry is zero no matter the input carry. If the two digits are one, then the output carry i…
comment
Yes, according to Apple official documentation, OpenCL was deprecated as of macOS 10.14. It is reported to work, including on Apple Silicon (M1 and M2), but don't expect any updates. [1]: https:…
comment
These chipsandcheese articles are fantastic, they really go into depth. Last week I had occasion to dig deeply into the microarchitecture of the Apple M1 and M2 GPUs (which are very similar to the A s…
comment
It's an instance of "argument from fallacy." This whole comment section is like a zoo of logical fallacies, with golden oldies such as ad hominem (both sides), false equivalence, and of…
comment
I'm not actually salty. Rather, sometimes I half-joke that I have 28 years of Rust programming experience, and when people do the double-take I explain that I'm including ML-derived language…
comment
My one criticism of this post is referring to the work of Aiken in improving the static memory management in ML Kit. That should be "Aiken et al".
comment
The standard TeX algorithm doesn't penalize the length of the last line, so doesn't solve this. The implementation in Android, based on Minikin[1], applies some penalty to the last line by…
comment
I wouldn't describe a GPU architecture as VLIW-like, though there is some overlap. To me, the essence of VLIW is extracting instruction level parallelism by specifying lots of different operation…
comment
Very nice story. Something like this would appeal to me, actually. It reminds me of Nesin Mathematics Village in Turkey. There are also some similarities to Recurse Center, where I did do a 3-month ba…
comment
A lot of the cultural phenomena around Rust can best be explained in religious terms. It's no accident that we are sometimes referred to as the "Rust Evangelism Strike Force," and the w…
comment
Going through the comments provides another datapoint for what I have observed as an ironclad law: the first version of any code involving tricky atomics and synchronization is always wrong. Reasoning…
comment
That is correct, here is a playground link which shows that result: https://play.rust-lang.org/?version=stable&mode=debug&editio... …
comment
I can speak to the UI question, as I'm very invested in it. It is way too early to standardize any such thing now. There are, give or take, 5 promising UI toolkits, and a dozen others that are …
comment
This one is fun! However, that particular formula is not my favorite, as it doesn't behave nicely when a is small. The art here is deciding what to special case. The way I approach this[1] is to …
comment
Previously: https://news.ycombinator.com/item?id=16312317 My favorite line from the paper is "the popular and elegant X86 architecture". Very subtle humor, I suspect not man…
comment
I have an exception that proves the rule. I thought about responding to Julia's call, but decided this was too subtle. But here we go... A central primitive in 2D computational geometry is the or…
comment
Another real world example, and actually quite relevant to the use case of the original article. Tint and naga are both shader compilers, primarily used for compiling WGSL into SPIR-V, MSL, and HLSL s…