probablybanned

The iPhone 7 is obviously inferior to the 3GS. For one thing it's larger, which is ridiculous, I mean I thought the whole point was to make phones smaller. And, can you believe it, it doesn't even have a headphone jack! Obviously worthless, throw that thing in the trash!

That's what I hear when people bandwagon bash the F35 without reference to avionics, comms, sensor systems, or low observability features.

In part, people are discontented that for the cost it's not as clearly superior in every way to the 4th generation as the 4th was to the 3rd. That's really because the manned jet fighter, like the smartphone or main battle tank, is a maturing technology.

The real frontiers of military innovation are in autonomous weapon systems and space warfare. So be careful what you wish for, I guess.

Supplemental: Listen to this and tell me how enthusiastic these "moderates" are about a democratic future for Syria. At this point you cannot convince me that they are anything but utopian Islamists who want to manipulate the United States into killing for them. If we dispensed with Assad, they would immediately get down to the business of fighting each other.

https://youtu.be/e4phB-_pXDM?t=1548

I have no patience for people who cry crocodile tears for a supposed loss of civility and use it as cover to push utterly cynical policies that will have massive real world impact. It's disgusting, actually, and not so different from the crybully Left.

Sorry for the belated response.

https://wikileaks.org/clinton-emails/emailid/23225

The "it's a big mess" excuse is just a fig leaf to cover the functional US policy in Syria, which has been to support Jihadists against Assad.

Yeah absolutely, they're called negligent discharges for a reason and I wouldn't personally hesitate to buy a Glock.

However, there is a distinction in that the essentially safetyless, essentially SAO trigger system is fundamental to the performance of the gun's function, whereas the takedown procedure is more of a peculiar wart of the design. So those idiots end up shooting their walls (and potentially other people) for no reason other than that Gaston Glock didn't want to compromise on the parts count. Typical engineer, I suppose.

After over a hundred years of iteration on the semiautomatic combat pistol, the top tier offerings from the respected manufacturers are all equivalent and fit for purpose. They pass the stress tests. The various pistols I've owned from Sig, CZ, Walther have all gotten up into the low thousands of rounds with zero malfunctions. Boring!

The only interesting development has been the convergence of all the major players on the Glock-style striker fired system, due mainly I think to the consistent first-round trigger pull which is a big deal for practical shooting, and a cultural shift towards accepting the... shall we say, less idiot-proof design. Although, everyone except Glock seems to agree that it's a bad idea to require a trigger pull as part of the takedown procedure.

e: And the "low thousands" is because I got tired of spending so much money on ammo

Yeah, instead of writing a small tool to solve your own problem quickly you ought to spend your time pushing this giant boulder up a hill. Good luck with that pitch!

I would say the same things if they were merely pitching a fit about being on the losing side of a free and fair election!

You've got one thing right, which is that not all of the outrage is feigned, and some peoples view of the world is so distorted that they would place their "struggle" under Western democracies with that of the some of the most oppressive societies on Earth.

Of course they're completely mistaken, and you try to cover for this cognitive failure with this feeble argument from "life experience" which can be used to justify nearly any kind of ridiculous behavior. Think about it: Our experience of the world is fundamentally subjective, so who are we to say that one way of acting is better than another? Fine, enjoy your postmodern intellectual paralysis if you want but I'm not playing that game.

The facts are what they are, and Donald J Trump may be a formidable political opponent but he's not the god damn Chinese Communist Party.

I'm not a nuclear engineer so I hesitate to shit up the thread with my armchair opinion, but I'm a big fan of small, low maintenance reactor designs like the Toshiba 4S. I feel like that's the way it ought to go: Away from large stationary reactor installations that require a huge amount of skilled labor on site for construction and maintenance, and toward modular banks of reactors that can be taken offline and refurbished off-site.

With a smaller, one size fits all design, one might expect to find economies of scale that don't currently exist. In much the same way, SpaceX seeks to make Falcon launches routine and push down costs of their incredibly complex machine through rapid iteration and refinement -- although reactor failures in the field must not be of the "energetic" variety of course.

Are there limitations of physics that will keep nuclear power stuck in the mainframe era, or might we see a commodity revolution?

It depends on the definition of negligence, of course. Sometimes you take all appropriate precautions and still get screwed -- OK. We have a situation where companies are releasing products that every competent professional would recognize as a security trainwreck, and it happens with complete regularity, because the incentives simply do not currently align to make corporate executives give a damn.

Due to externalized costs that impact more or less the entire population, this is now a political issue. It's exactly as thorny as environmental regulation, or public safety regulation. Should users truly own their cars and have the ability to disable safety features? I think so, but it shouldn't just be a button on the dash. Should they be able to disable emissions features?

I'm not trying to oversimplify here, but I also don't think we should allow the perfect to be the enemy of the good. If there were real incentives to produce secure products, we could expect to see much more investment in secure software, perhaps even verifiable formal methods, and hopefully more industry collaboration/standardization around open source platforms to mitigate risk. It's a tradeoff, but I so often see front-line engineers faced with situations where acting responsibly/ethically with respect to security puts them at odds with management, and that's a clear sign that we're not getting the balance right.

I already see your comment turning gray for daring to suggest a role for regulation here. Sorry people, if your company releases a negligently insecure IoT device into the world in any sort of quantity, you are a polluter.

The latter set of annoyances are of course what Apple are intending to solve with the W1/AirPods -- intelligent switching between all iCloud devices without re-pairing.

I personally favor Bluetooth headphones even with all the annoyance because as bad as it is, cords are even worse.

For those who feel differently, I suspect they'll either use the bundled earpods or just leave the lightning adapter attached to their best pair of headphones and move on with their lives. Maybe I'm wrong and everyone except me is rotating five pairs of headphones between six different devices on a regular basis, but it seems like just another case of piling on to resist change.

This is sometimes the case, but the most popular cards on the market are indeed single chip ASIC solutions.

Also, don't forget that small FPGAs like those from Lattice Semi are very cheap these days and find a lot of use specifically in obscure IO interfacing tasks.

Hey, I picked this username for a reason.

Yeah, I would argue that both of those functions are re-entrant. The question is whether you can have two calls to is_not_reentrant executing concurrently without interfering with each other. So a typical example would have a non-reentrant function altering a variable like global_int as part of its computation.

Also, it's important to think about the kinds of program transformations that the C compiler is free to do with your code when going from C to assembly. I think it's likely that your compiler would produce the same output for both functions. Lines of C code do not execute atomically, and instructions can be re-ordered.

EDIT: Returning to this because I think I made a mistake. "Two calls executing concurrently" is not exactly right, because it admits the possibility of arbitrary instruction interleaving. A reentrant function must be able to be paused at any point, have another call to the function execute entirely, and then resume correctly. This is a weaker condition.

That's a vacuous statement. She said "big" gaps. And I'm saying that she's making errors that no one with a background in CS (e.g. a degree) ought to make.

I have no idea what your agenda is in lavishing all this unnecessary praise on Morgan Phillips, but it's gross. Please stop.

Really? Go take a look at the two code snippets in "The Reentrant Kernel." I can't tell what the salient difference is supposed to be between them, much less how they are meant to illustrate the concept of reentrancy.

I mean, I appreciate the concept of what she's doing here, but that self-assessment sounds reasonable to me.

For me, AOT compilation is not about shipping binaries vs. source. The software I write is mostly used for internal research, so it's not going anywhere. It's difference between getting compilation errors from 'make' or having to pull them out of the OpenCL C API at runtime. Setting a breakpoint in an OpenCL kernel... I remember Intel's stuff basically working fine, but you had to pass a vendor-specific option to hint at the file path. In all these little ways, the workflow is just sucky. Much of it can be worked around, but I'm too lazy to write more application code to do all the chores that the CUDA toolchain takes care of already. I'm glad to hear that 2.0 is fixing this.

What's your alternative to templates for generic code, exactly? C macros? Scripted pre-processing that further screws up the already marginal tool support for debugging and profiling? Copy+paste? Templates are completely orthogonal to "complex computation" -- I just want to use a device function on different data types without run-time overhead.

On the topic of complex computation, I'm constantly surprised at the kind of features NV adds to CUDA and how well they actually work. I'm also surprised at the kinds of things people do on the hardware. If someone implements a high performance lock-free data structure on the GPU, you can't look at that and say oh, that's too complex, you shouldn't do that.

Also, until we're all working on computers that look like the PS4 with a unified global memory, there's a huge incentive to cram the awkward bits of your program onto the GPU any way you can, even if it drags a bit, because that's where the data is.

NVIDIA has really gone to some extremes. malloc in device functions. vtable support. Dynamic parallelism. Metal has none of this stuff.

<EDIT: Just now saw your other replies downthread. I'm leaving this comment because it reflects my personal experience and opinions, but don't feel like you need to repeat yourself to clarify your position re: templates, etc>

I agree with you. I'll bet this is coming to OS X sooner rather than later, and people are largely missing its significance as a GPGPU programming environment.

Despite Apple's best efforts, OpenCL uptake seems to be sluggish. CUDA continues to dominate developer mindshare, by providing a far better language, API, and toolchain.

Compare the C++11 subset supported by the Metal shading language to the device language of CUDA C++. Templates are a huge feature. Ahead-of-time compilation is huge (vs shipping strings to the driver like OpenCL). It retains the basic workgroup structure of OpenCL with local and global memory, so it looks feasible to map to NVIDIA and AMD hardware. Is there really anything PowerVR specific in here? People seem to be inferring an awful lot from the name, but nothing sticks out at first glance.

The features of the shading language would make porting applications from CUDA less painful. If they went all in on XCode dev tools to make it a rival to NSight for profiling/debugging, maybe those Mac Pro GPUs wouldn't seem so neglected.