HNHacker News
TopNewBestAskShowJobs

merijnv

1,071 karma · joined January 10, 2010

merijn@inconsistent.nl
submissionscomments
merijnv··on Does not GPU compute
> Given the quality of OpenCL and its cross platform nature

I'm sorry, WHAT? OpenCL is absolute shit. Cumbersome API definition, lack of low-level control, stringly typed programs (all programs are provided as strings and kernels are identified with those too). Which means nearly no compile-time feedback, it's hard to embed GPU kernels into a single binary. The API is woefully lacking in flexibility (no dynamic launch), OpenCL 2.0 is better, (EDIT: Apparently AMD supports it now, I'd have to check whether Intel/NVidia have also added support), but no one supports it so it's also irrelevant.

Not only that, AMD hardware is terrible. Atomics on NVidia's maxwell are orders of magnitude faster than on AMD (to the point of being comparable to non-atomic operations with low contention).

CUDA's environment provides: Better documentation, better feature support, saner development and debugging, possibility to ship both generic & specialised binary kernels, JITtable kernels in intermediate representation, better compile time sanity checking, the ability to generate your own IR/CUDA assembler from non CUDA languages...

The reason everyone does CUDA and uses NVidia is because there's zero real competition. AMD is the only company that cares about OpenCL, Intel and NVidia just implement the bare minimum to have AMD's OpenCL code be portable to them. Intel has OpenMP and TBB for the Phi, NVidia has CUDA.

To me it's crazy that anyone keeps mentioning OpenCL as a serious alternative. In theory I agree that an open standard would be nice, but over here in reality where I have to actually write code there is no realistic alternative to CUDA if you want to stay sane.

merijnv··on All European scientific publicly funded articles to be freely accessible by 2020
Ahahahahaha.

First of all: Researchers won't avoid public funds, because there's never enough money.

Second of all: Most researchers HATE academic publishers and being forced to sign over their copyright to a paywall in exchange for nothing. But institutional evaluation criteria force you to do so.

EU banning non-open publications will force publishers to allow open access OR force universities to revise their criteria to accommodate the inability to publish publicly funded work in non-open journals.

I think all of my colleagues think this is a great thing. As do I. Elsevier can go blow a goat with their copyright assignment BS.

merijnv··on Fitbit Acquires Coin
Is US banking really this dystopian?

I've had my Maestro debit card skimmed twice, causing about 3k to be withdrawn from my account. In both cases my bank had:

* Filed a police report

* Disabled the card

* Mailed me a new card

* Started recovering the money

Before I even noticed the money being gone. The first time I actually had my money and new card back before I noticed my old card got disabled. The second time I noticed because I couldn't buy lunch using my card, but in both cases I had my money and a new card within like a week?

Besides, no one here really has/uses credit cards and stores mostly don't accept them.

merijnv··on Academics Make Theoretical Breakthrough in Random Number Generation
Sure, if you're hardware RNG is inspecting your RAM and trying to compromise your entropy pool that could be done, but:

1 - The entropy mixing is more complex than simply XOR, making such a thing considerably harder

2 - If you expect this level of backdooring from your CPU, you have bigger problems :)

merijnv··on Academics Make Theoretical Breakthrough in Random Number Generation
Linus actually argued something slightly different. That discussion was about whether a "bad" random source can weaken a "good" source. This paper is about constructing a "good" source from two independent "bad" sources.

The problem with Linux was people complaining that mixing the Intel RNG into the existing entropy pool would lower the entropy of the entire pool, effectively letting a backdoored Intel RNG render the random system untrustworthy. Linus' response was the fairly simple/logical observation that "that's not how entropy works".

Suppose we have a random bit string with entropy X and we xor the bit string with a completely deterministic bitstring, what's the result? Clearly you have a bitstring that STILL has entropy X, because the result is completely dependent on our initial bitstring. Mixing with a deterministic bitstring doesn't make the output of our XOR any more predictable then the original bitstring, it's equally random.

So Linus' argument was that "IF the Intel RNG is completely backdoored and deterministic mixing it with the entropy pool will have no effect on the entropy in the pool. HOWEVER, if the Intel RNG is anything but completely deterministic, i.e. there is even a tiny bit of randomness in it, this will actually INCREASE the pool's entropy.

So mixing a completely backdoored RNG will have no negative impact, but mixing anything that's not 100% predictable will have a positive impact, so there's no reason to not always mix the hardware RNG with the pool.

merijnv··on Why Apple Music is So Bad When the iPhone is so Good
The Spotify "Discover Weekly" seems to do pretty well at finding stuff I like. Although it does seem to be a bit biased toward some of my tastes more than others.
merijnv··on Email privacy bill unanimously passes U.S. House
You appear to be ignoring the fact that even if you consider the NSA's work benevolent, it is only benevolent to US citizens. Meanwhile lots of us on here are, in fact, not US citizens, especially at this time of day. So the NSA's activities are all negative, no positive to non-Americans.
merijnv··on Once upon a time, memory allocators made sense
> (a) malloc() never returns NULL. It always returns a valid address, even though your system may have out of memory.

This is only true on Linux, for example, FreeBSD by default has partial overcommit and WILL start returning null after overcommitting some percentage of memory (I think 30% or so by default)

merijnv··on Swift is Open Source
Unspecified, so legally murky. But there's a notion of "implied license", so someone using the code could argue that "openly putting it somewhere publicly" constitutes and implicit license grant.

Note by the way that you are ALREADY explicitly granting some rights by putting it on GitHub:

We claim no intellectual property rights over the material you provide to the Service. Your profile and materials uploaded remain yours. However, by setting your pages to be viewed publicly, you agree to allow others to view your Content. By setting your repositories to be viewed publicly, you agree to allow others to view and fork your repositories.

Arguing that this constitutes an implicit license to use this code probably wouldn't be too hard. However trying to relicense/redistribute is less likely. In the end this will have to be decided by a judge in court.

merijnv··on Web login using SSH
So...basically you want SSL client certificates but with a slightly less sucky UI?

I mean, https already supports this, it's just that setting up servers to deal with client certs sucks and the UI for users is also pretty bad. But the infrastructure is already there without poorly reinventing the wheel over SSH.

merijnv··on FastMail: Shutting down our XMPP chat service
> and the time that Google stopped federating with other XMPP server operators.

I see this keep getting repeated, but AFAIK there was only like a month long time they temporarily suspended federation for adding new contacts due to a spam issue? I've been talking to people on Google Talk from my (FastMail, sadly) XMPP account for the past year without issue? Am I going crazy?

merijnv··on FastMail under DDoS Attack
I've been using Fastmail for 2 or 3 years, recently moved my XMPP and Caldav stuff over too. I would recommend them, no complaints, I've not noticed a single outage in that time.

It was trivial to IMAP import mail from my old host. Good support for multiple authentication schemes, i.e. I have a 32 random char master password saved in my password manager, and they allow me to add any number of alternative authentications like shorter passwords that are only allowed in combination with Yubikey/Google Authenticator or one time passwords.

As for "app" I just use a generic IMAP client.

merijnv··on The Theory of patches-vector
Yes! http://bartoszmilewski.com/2014/10/28/category-theory-for-pr...

It uses both Haskell and C++ for examples, although for best effect it'd help if you could read (basic) haskell types. Or maybe not, I read it after learning haskell so I find it hard to just how it is from a just C++ perspective.

merijnv··on Global coalition tells Facebook to kill its Real Names policy
> New legislation is a direct threat to companies, backed by the power of the government to fine and arrest those who disobey. Why go to such lengths when it's much easier just to stop going to Facebook? Only a perverse sadist would prefer legislation over ostracism.

Right, this is why the US is such a shiny beacon of safe products, consumer-friendliness, and consumer rights compared to the horrible European markets... :)

I always find this sentiment on HN so amusing. "Legislation hurts companies, we should all just ostracise/boycott companies instead!" despite all practical evidence suggesting that legislation actually gets things done. Like, I can actually get fast affordable internet here, unlike when I was in the US :)

merijnv··on Global coalition tells Facebook to kill its Real Names policy
> the only way to force companies is by spending your money somewhere else

Or, you know, enforcing new legislation that forces the company to do what you want...

merijnv··on Haskell's Type Classes: We Can Do Better
The origin of a name is statically known in the location (i.e. file, expression, whatever). The origin of data is dynamic and can vary at runtime based on arbitrary previous evaluations. Unless you're going to propose making types unique (i.e., not compatible) depending on their origin and enforce never mixing them in the compiler.

At which point you have basically reinvented "newtype"

merijnv··on Haskell's Type Classes: We Can Do Better
But then my ability to insert elements into a Set depends on the location where those elements were created...that sounds awful.
merijnv··on Spotify Just Got Real Creepy with the Data It Collects on You
To be fair, I opened the privacy policy from Spotify just now (I think I updated to a newer version yesterday?) and I don't see any of these things mentioned in the privacy policy, so might wanna hold your horses.
merijnv··on Spotify Just Got Real Creepy with the Data It Collects on You
FYI, I opened up the privacy policy from my Spotify install (I updated yesterday, I think?) and none of this is mentioned in it.
merijnv··on Swift Functors, Applicatives, and Monads in Pictures
To be fair, I don't know a single person that explains things using

> "just a monoid in the category of endofunctors"

In fact, even among people who DO know the category theory necessary to understand this expression it's considered a very obtuse and useless explanation. Especially in the context of Haskell.

As for good readable introductions, I consider Wadler's paper "Monads for Functional Programming" to actually be a pretty good introduction. Another good intro is "You Could Have Invented Monads! (And Maybe You Already Have.)" http://blog.sigfpe.com/2006/08/you-could-have-invented-monad...

merijnv··on A Dutch city is giving money away to test the basic income theory
> Because there is no free money.

Ok, sure.

> We would all have to pay for it.

Well, yes.

> We would be subsidising laziness.

Why is it laziness? As I pointed out, repeatedly, above. There's two ways we can go with this discussion. You can state "There is nothing worth of value, except economic value". If that's what you believe, fine. But I think that is a very shallow and hollow belief and then there is not much sense in further discussion.

Alternatively, you can agree that there are pursuits that, while not producing any economic value, are very valuable to society. Such as arts, social projects, research, etc. and that those of us who are good at maximising economic value (but, most likely not spending as much time/effort in other areas) should subsidise those of us who ARE taking care of our elderly, raising kids, or otherwise fulfilling socially and artistically important roles. In which case we can discuss whether basic income is a good way to accomplish this or whether there are better ways.

And what is so terrible about subsiding laziness? Why do you care? If you have more than enough money to do whatever you want, what does it hurt you that someone manages to "not starve" even though they don't "do anything"? Is that really so bad? You would honestly prefer thousands, if not millions, starving to avoid "subsidising laziness"?

merijnv··on A Dutch city is giving money away to test the basic income theory
> But no - that doesn't mean the state should give everyone free money.

Why not?

merijnv··on A Dutch city is giving money away to test the basic income theory
Struggling everyone. Why the hell is some bank CEO making several hundred million a year and we have people that can't even reliably be home in the evening to tuck their kids in because they're working 5 jobs. I'm all for taxing high-end income (coporate and private) even more to pay for this.

Why would everyone need to work anyway? It's not tenable. A few more years/decades and we'll have automated so much it'll be impossible to actually employ everyone. So unless you suggest we let everyone who is in a now obsolete field starve, I suggest we start thinking of how we're going to manage a society where it's just impossible for everyone to work.

EDIT: Also, where's this 240 billion figure coming from?

merijnv··on A Dutch city is giving money away to test the basic income theory
Who says it has to be a company? How about that guy that wants make music but doesn't have the time to practice due to work. The girl that wants to paint, but is to tired after work.

Why does everything have to be about companies and money? Can't you really imagine any useful social, artistic, or other things these "dumb, lazy" people might want to invest their time into? How about just being a stay at home parent and investing more time in their kids? Do you really think society would not improve from any of these things?

merijnv··on From Haskell to Hardware [video]
> "Don't be fooled, though: Even in Haskell you can write very ugly code."

> Also very much the case - I've done it!

On the bright side, refactoring ugly code into not ugly code is substantially easier in Haskell than many other languages! ;)

merijnv··on Why Greet Apple's Swift 2.0 With Open Arms?
Short summary (from memory, so might be a little faulty):

GPLv3 = GPLv2 + you can't sue any users/distributors for violating your patents + you grant an implicit license to all your relevant patents to your users + you must protect users from patents that might be covering code you contribute

For companies with patents and lots of patents they are afraid of losing their ability to sue for patent violations. I'm unclear whether the license you grant is just for the GPLed code or just in general. And in general it's unknown how all this plays out in court, so most lawyers (and thus companies) avoid v3 like the plague.

AGPL = GPL + must distribute even if you just provide an internet accessible service.

The idea being that the GPL only applies to distribution, so if I modify some GPL software but never distribute it (because I just run it on a server I control) I never have to GPL my changes. The AGPL says you have to distribute/open source your changes even if you don't distribute if the modified software is accessible over the network to users.

merijnv··on Tarsnap GUI for the desktop
Where are you getting this from? The latest release on the website is from February 2015, as near as I can tell. Or at least, that's when it was signed.
merijnv··on Stack – a Haskell development tool
Eh...xmonad and pandoc are pretty popular projects, both written in haskell
merijnv··on Cloud Haskell
FYI, sparks are different from threads. Haskell has "forkIO" threads which are comparable to Erlang processes/go-routines (actually, they're cheaper than both since haskell allocates less stack for threads). Threads are what you want for concurrent programming (i.e., non-deterministic interleaving of behaviour, which with GHC can be run in either single-threaded or multi-threaded mode).

Sparks are for deterministic parallelism, the idea behind sparks is that you create a pool of sparks, each of which represent a small, finite bit of work that can be done in parallel.

In general sparks will represent a finite, deterministic amount of work, whereas threads represent a (potentially) infinite amount of non-deterministic work.

merijnv··on Vim's speed is not really the point (2013)
The problem is that the main use I thought to have for it is jumping to functions, but I usually do that using / (search) and/or ctags (Ctrl-])
← PreviousPage 4 of 11Next →