HNHacker News
TopNewBestAskShowJobs

catpolice

666 karma · joined July 23, 2015

submissionscomments
catpolice··on Is e Normal?
"Is it conceivable that different irrational numbers have different "frequencies", such that high-frequency numbers exhibit their expected averages very quickly, whereas low-frequency numbers may require billions of digits to exhibit the expected averages?"

Yes, and obviously so. Take a (base 10) normal irrational and divide it by 10^1000000000000. The first trillion or so digits of the result will be 0, but the result is still irrational (because duh) and (base 10) normal because the frequencies of the digits have to converge in the infinite limit and giving one digit a large finite head start won't change its overall proportion in the limit. QED.

catpolice··on Tensorflow 0.6.0 Release
Cool, thanks!
catpolice··on Tensorflow 0.6.0 Release
Does this version support the latest CUDA/CUDNN? I spent hours with the initial release trying to get GPU acceleration working before giving up in frustration (and being a seasoned Arch linux user, I'm not easily discouraged) - I can't remember what ultimately caused me to throw my hands up in defeat but it had to do with the old CUDA version requirement.
catpolice··on D3.compose – compose reusable charts and components with D3.js
Hmm, here's my experience:

What I've needed from D3 is the math-y bits: calculating layouts, paths, etc. What I don't need (or want) is for it to ever touch the DOM.

But that's what it wants to do most. D3 is sort of the best of the jQuery-inspired "select and mutate" generation - it does data binding in a way that really makes sense as long as you let it handle DOM operations.

But following React, there's been a pretty big movement away from procedural DOM manipulation and toward the declarative re-rendery "DOM as a pure function of state" approach. In that case, letting D3 mess with the DOM (especially in the way that it does, where complex links between DOM and data are maintained in the background) is bad news.

If you already have a way of getting elements in and out of the DOM and keeping their state aligned with the data that you like better than the select-and-mutate model, a lot of what D3 does seems unnecessary or introduces problematic interference. React and D3, for instance, tend to not play very nice without a lot of effort.

I've had projects where programmers lean on D3 because that's the way they've learned to get SVG elements into the DOM - they'll take an example they've found somewhere that mostly only uses simple shapes, modify it a little bit to do what they want and ask me for help adding that visualization into another project. Sometimes in these cases, D3 isn't getting used for much except for "insert an svg element with these attributes for every node in this data structure", in which case it's total overkill, and I end up stripping it out entirely. I sometimes end up keeping it around for things (I'm probably not going to write my own Reingold-Tilford tree layout algorithm), but often in these cases I have to do extra work to keep it from conflicting with other things.

catpolice··on A Story about Symbolics Lisp Machines
Texas yes, College Station no.
catpolice··on The sky's gone dark
There's also a really, really large amount of space for maneuvering up there. Keep in mind the surface area of the sphere that's at 36,000 km from the ground (approx. average satellite orbit height) is ~22,567,880,697 km - that's entirely disregarding vertical room for maneuvering.
catpolice··on Thermodynamical cost of some interpretations of quantum theory
For various reasons, in quantum mechanics, one typically excludes the observer from the system being modeled. This is obviously problematic, especially if you're doing large scale cosmology. But that's the measurement problem for you.

Regarding the context, the quote I'm thinking of comes from one of the essays in Speakable and Unspeakable in Quantum Mechanics, which is an enlightening read, both in terms of content and for its historical value. The issue, I think, is that the kind of 'conspiracy' implied by superdeterminist theories, wherein nature sort of guides our hand in picking certain experimental settings so as to hide information seems troubling to many scientists. See Zeilinger's quote at https://en.wikipedia.org/wiki/Superdeterminism

catpolice··on A Story about Symbolics Lisp Machines
We use one of these in the lobby as a coffee table at my work.
catpolice··on Thermodynamical cost of some interpretations of quantum theory
"Independently" is also a loaded word - one could mean that there's no direct causal connection between the state of the system and the state of the measurement apparatus, or one could mean that those two states are (in ensemble experiments) independent in the statistical sense.

In order for Bell's theorem (and a lot of subsequent quantum theory) to work, we have to interpret causal and statistical independence as being one and the same. But the more one learns about the theoretical justification for this conflation, the shakier it seems. Bell, in particular, justified it by claiming that it was unscientific to assume that experimenters didn't have free will. Okay, sure.

Any universe where the assumption that experimental measurement settings are always already statistically correlated with the state of the system being measured (even if that correlation would have to have been established millennia ago) is fundamentally weird in certain ways - these kinds of theories are sometimes called 'conspiracy theories' as it sort of seems like all the information is there and nature is conspiring to hide it from us. But there's a whole branch of super-determinist interpretations of quantum mechanics motivated by the sense that this kind of weirdness is not as bad as the kind of weirdness we'd have to otherwise accept. This kind of work has been out of fashion for a while, but it seems to be gaining a certain amount of momentum in the last decade or so.

catpolice··on Procedural Dungeon Generation Algorithm
I've been working on an algorithm to make more playable dungeons with included puzzles (e.g. locked door/key quests etc.). The math is there to ensure everything is solvable and relatively interesting (you're guaranteed to run into a puzzle before you find all the elements you need to solve it, for instance) but I got bored when it came time to make an example game to illustrate how it works. I'll pick that back up at some point...
catpolice··on React, JSX and ES6: The Weird Parts
Absolutely, but at the same time, sacrificing client side rendering performance isn't the answer, and some of his suggestions about how he wishes React worked would involve doing just that. The author's complaint isn't that server-side rendering is harmfully impacted, it's that JSX has some expressively awkward bits. But it has some of those awkward bits because it more or less has to if React is going to work as well as it does on the client side. JSX was designed the way that it was to enable client-side functionality, and the author doesn't seem to acknowledge that, so it seems like he's criticizing without fully understanding.
catpolice··on React, JSX and ES6: The Weird Parts
Yeah the author seems to be thinking about the DOM from the server side, where the end point of all rendering calls is a string that gets sent to the client. This misses out on baaasically everything that sets React apart from all the template-based frameworks out there.
catpolice··on Ask HN: Should I use Polymer for my next project?
I'm optimistic about web components, but I don't think Polymer's ready yet - I was seeing their demo components not work in a new Chrome on Linux just a couple months ago (they still may not work, I'm too lazy to go through the examples right now).

The main two draws of Polymer are that it's built on web components (which are cool and futuristic) and that it implements material design (which is nice looking). There are other material design widget libraries however, even ones made by Google. Have you seen MDL (http://www.getmdl.io/)? It's a more minimalistic CSS/Javascript approach and will probably be faster and more cross compatible.

FWIW the three things you mentioned (Polymer, Angular, React) are kind of dissimilar in an apple-to-oranges way:

* Polymer is basically a UI widget library with some other bits thrown in - it doesn't really help you structure your application, it's just widgets for building the view layer.

* Angular is a full-ish MVC platform that provides you with tools to build your own UI components/widgets

* React is just a toolkit for building UI components, and it doesn't come with anything else (but it does encourage a kind of design where you don't need much else)

Neither of the latter two come with pre-built widgets: you have to make them yourself or find libraries of them made by others (which is pretty easy to do).

For my money, your expectation on which of these is easy to pick up is backwards: Angular is way, way more complex than React and has a much higher learning curve. React does basically one thing (and does it very well) so there's really not much to learn about it at all - you can pick it up in a day once you wrap your head around the rendering-as-a-pure-function-of-state concept. Angular is a full framework and it does a lot of things and is opinionated about how it does them, so there's a whole lot more to learn. If you need all that MVC stuff done for you, Angular might be the way to go though (but beware the coming 2.0 transition that's going to break all backwards compatibility completely).

I personally like React a lot - I'm probably going to try to wrap MDL components in React for a project I'm working on right now.

← PreviousPage 5 of 5