On the other hand, pretty easy how such a div might trigger a less efficient path; if the video is top in the z-order then it can probably bypass being composited by the browser (and who knows, maybe even bypass being composited by the OS) and avoid a whole mess of rendering to a texture, texturing some triangles, and so on and so forth.
Right. My take is "hey, good engineering to not need a power brick. But to what end? What's the practical purpose of this engineering? Why not just make the system bigger?"
One of the (many) reasons I've heard they had to abandon Longhorn was that apparently it was essentially unbuildable. The build was broken so much of the time that they gave up and started over.
Even with the CPU, you also need the right chipset and the right firmware to actually light this stuff up. While especially in the laptop sector there are consumer devices that include this, it's far from universal.
Good luck with that. It's working for Raytheon with a clearance.
> (C) discrimination because of citizenship status which is otherwise required in order to comply with law, regulation, or executive order, or required by Federal, State, or local government contract, or which the Attorney General determines to be essential for an employer to do business with an agency or department of the Federal, State, or local government.
It isn't on all hardware. Intel has two ME firmwares, a small one for consumer systems, and a big one for corporate/enterprise systems. The small one does not (or at least, should not; is not supposed to) include the remote management features.
In other words, the separation that you describe exists.
Systems with the full firmware sport things such as the vPro branding, and only certain combinations of CPU and chipset support it.
Sure, if you install a GNU stack (glibc, for example, being an important part) you can run typical Linux software on Android. It's not the default, however. While the Linux kernel has proliferated, the userlands are more varied; Android/Bionic on phones and tablets, uClibc on routers and other embedded roles, GNU/Linux for most servers and desktops--though even in that final category, we see substantial variation such as systemd versus init.
Sometimes this is to Linux's advantage; it underscores its flexibility, and the ability to use, for example, busybox uClibc on small systems is definitely an advantage. But other times, like having no standard UI toolkit that works across desktop, tablet, and phone, or having different approaches to managing services/daemons, the advantages are less clear.
Canonical was heading hard in this unified direction, but I'm not sure what their current phone/tablet plans are. Their stupidly implausible kickstarter seems to have disrupted these efforts.
Modern graphics hardware isn't optimized for drawing bitmapped lines. It's optimized for drawing textured triangles. This means that it's not actually a good fit for GDI any more. Given that Windows Phone was, as far as apps were concerned, a clean slate, there was little reason to retain USER and GDI.
win32k.sys is also rife with security flaws, so scrapping it has arguably reliability benefits too.
> Snark aside, this is probably one of crappiest articles I've ever read (EDIT:) on Ars Technica (or rather, skimmed; didn't deem it worth reading word-for-word). "Microsoft should make a do-it-all box for $200 that has cutting-edge GPU and CPU, and they should fragment their hardware lineup. Oh, and put a pony in every box." I'm sure this article is being passed around the E&D executive suite as I type.
Maybe you should bother to read the article before commenting, because it says quite the opposite, and suggested that in fact it's the attempt to make a do-it-all box that leaves the Xbox simultaneously overpriced as a streaming box, and underpowered as a console.
HoloLens is a derivative of Kinect. Kinect 1 (with the Xbox 360) was structured light. Kinect 2 (with the Xbox One) is time-of-flight. I'm not sure which tech the HoloLens uses (I don't think Microsoft has specified one way or another, and I didn't have an IR camera to look to see if it was emitting a pattern), but I'm assuming it's based on the newer tech, because my understanding is that time-of-flight systems can much better handle having multiple devices scanning the same space. HoloLens handled this situation very well; I've seen at least 8 devices all scanning the same area without getting confused.
It's all driven by the unit, btw; no need for markers or external illumination.
Conker can handle complex objects, but they have to have been present when scanning the scene; my cameraperson was moving around so she didn't get scanned.
The depth maps can get dynamically updated, so it's not insurmountable, but currently the process seems to take a few seconds. I'm not sure if this is a design decision, a processing limitation, a hardware data capture limitation, or something else entirely.