Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
mwkaufma
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
21 ms
·
421.
▲
by
mwkaufma
6y ago
Skipping the self-answering rhetorical question. It's weird to claim that a studio which flagrantly plagiarizes "also has some redeeming original work." It's not like there's this clear line in the creative process
422.
▲
by
mwkaufma
6y ago
So FASA is a victim here because some of their designs were original derivatives, and not straight-up stolen from Macross, Crusher Joe, and Dougram?
423.
▲
by
mwkaufma
6y ago
FASA is as shameful about never crediting creators as HG. Kawamori stated in an interview that they overextended the "package illustration" rights that they bought to Studio Nue mecha: https://www.forbes.com/sites&
424.
▲
by
mwkaufma
6y ago
Pet peeve: the primer includes Robotech/Harmony Gold (ip squatters) rather than Macross/Studio Nue (actual pioneers). The Macross Design Works series are much better artbooks, and actually contains production notes because they _a
425.
▲
by
mwkaufma
6y ago
Problem: too many standards. Solution: add another standard.
426.
▲
by
mwkaufma
6y ago
Love to see graphs without units of historical quantities that would be impossible to measure.
427.
▲
by
mwkaufma
6y ago
A critique from Bay Area planner Asher Kohn, also from 2017: https://www.asherjkohn.com/blog/2017/10/30/the-weak-logic-in...
428.
▲
by
mwkaufma
6y ago
Is the route count unbounded? O(n) = O(1) for constant-bounded n.
429.
▲
by
mwkaufma
6y ago
In my dayjob as a code-reviewer, I see it in code-bases a lot. Between the generic name and elevated status in the std namespace, it's a natural tool for developers to reach for, across experience-levels. I speculate that the boxing si
430.
▲
by
mwkaufma
6y ago
You are correct that the type the compiler creates for a lambda is allocated in-place, usually on the stack, and perform no heap-allocs. However if you pass it to a std:: function it will be _boxed_ and std::function _will_ heap alloc the s
431.
▲
by
mwkaufma
6y ago
I'm not saying that the default allocation strategy isn't good (in the general case), just that it's not customizable (for special needs).
432.
▲
by
mwkaufma
6y ago
Furthermore, C++ dependencies commonly instantiate types like std::vector with the default allocator internally, rather than exposing it to the host application.
433.
▲
by
mwkaufma
6y ago
Many core modern C++ types don't permit customizing the allocator. E.g. std::function
434.
▲
by
mwkaufma
6y ago
The author fails to mention the only truly-abusive us of this: working around readonly variables: struct Location { public readonly float x; public readonly float y; public Location(float ax, float ay) { x = ax;
435.
▲
by
mwkaufma
6y ago
Web-pack is not an elementary particle. It's a man-made construction. Describing its disk-size as an intrinsic aspect is a categorical mistake. Webpack is no-doubt a big program with a lot of features. But I have several even-bigger pr
436.
▲
by
mwkaufma
6y ago
Why can't it be 50mb? Why shouldn't it be 200mb?
437.
▲
by
mwkaufma
6y ago
"It's not frontend" isn't an answer. Why should a text-packager clock in at 100mb?
438.
▲
by
mwkaufma
7y ago
"... as a treat!"
439.
▲
by
mwkaufma
7y ago
Except it wasn't designed to be an "optimal target for SPIR-V" -- it was designed by Apple according to Metal's semantics, to promote Metal over Vulkan, and had to be rejiggered. It's worse, by your standard, than e
440.
▲
by
mwkaufma
7y ago
Lots of languages _already_ compile to SPIR-V - GLSL and HLSL for a start. Why add _another_ syntax?
441.
▲
by
mwkaufma
7y ago
The "why not UBI" section misses the most salient critique: what will landlords do when they learn all their tenants have an extra 2k in the pocket a month?
442.
▲
by
mwkaufma
7y ago
Good note!
443.
▲
by
mwkaufma
7y ago
It's not about syncing the arc, it's about syncing a result that is now _asynchronous_ (what did the projectile hit? How is the world changed by that hit? Could the projectile be dodged after it was fired?). For game-meaningful re
444.
▲
by
mwkaufma
7y ago
A library to magically solve the hardest fundamental problem of concurrency? Not so much.
445.
▲
by
mwkaufma
7y ago
Good primer, one quibble: ballistics aren't really computationally complex, as the article suggests, but rather _hard to sync over a network_. Local dead-reckoning to compensate for latency, but still ensure all players see the same re
446.
▲
by
mwkaufma
7y ago
XML vs JSON (vs code) devolves into bike-shedding because it positions the discussion in the wrong place. Developing layout in plaintext, rather than a WYSIWYG editor, is a priori impedance-mismatched.
447.
▲
by
mwkaufma
7y ago
"WinForms was just code" This conflates source text and product. Generated code was _output_ by the WinForms WYSIWYG editor, but the tool stored the layout source in a native project file format.
448.
▲
by
mwkaufma
7y ago
"We love services." Citation needed.
449.
▲
Insomniac's Web Tools: A Postmortem
(gdcvault.com)
3 points
by
mwkaufma
7y ago
|
0 comments
450.
▲
by
mwkaufma
7y ago
I would have loved to upgrade sooner, but I work in technical art and all our tools are version-locked by native python extensions and QT5's compat. See discussion at: https://vfxplatform.com/ I predict Maya devs (and,
More ›