Slashdot Mirror

So most likely Optium will now command the market, and the Laser projector will suffer the same fate as Betamax/Laser disk.

Well, it might not either - DLP is later technology than LCD, but both persist, and I'd even say DLP is gaining in projector fields.

The LED device is around the size of a wallet. The laser one prototype was the size of one of those small matchboxes. Going by this website, they've gotten that down to a penny. Going by Microvision's own site, they're looking to integrate the laser technology into a cell phone. If it can be made cheap, durable, and effective enough, it'll be able to carve it's own niche easily enough.

... But it has to be asked.

OpenBSD's developers (including OpenBSD 'benevolent dictator' Theo DeRaadt) have praised VIA Padlock functions in the past.

As a user of both OpenBSD and Gnu/Linux, I'd like to know if you share Linus Torvalds infamous appreciation of OpenBSD developpers? Or do you have good relations with all open source projects?

Wireless USB is a 128-carrier, OFDM system that instantaneously occupies 528MHz of spectrum. Each carrier is separated by about 4MHz, and in general uses simple modulation schemes to keep the complexity of the FFT processing to a minimum. While the theoretical speed is advertised at 480Mbps, practical demonstrations to date in real world environments, especially with more than one radio link running simultaneously in the same physical space, show more like 30-50Mbps. However, there's few real-world applications that truly use a significant portion of USB2.0's available bandwidth, so no one (except the external disk drive manufacturers) seem to be too worried about the bottleneck. WUSB was born from the FCC's ruling back in 2002 to open up 7.5GHz of spectrum, from 3.1 to 10.6GHz, to license-free, ultra-low-power, ultra wideband transmissions. The salient definition of UWB to the FCC was instantaneous occupancy of at least 500MHz of contiguous spectrum, thus the 128 carriers and 528MHz bandwidth. Otherwise, WUSB could be called an "802.11a on steroids", since it uses the same modulation scheme but with many more carriers. WUSB's promoters divided that spectrum up into 528MHz "channels", (see Figure 2 at http://www.deviceforge.com/articles/AT8171287040.h tml) and have a scheme whereby the link between two radios hops between channels in a psuedo-random fashion. Current technology radios use only the bottom three channels (3.1 - 4.8GHz) as current inexpensive CMOS technology doesn't perform well enough at higher frequencies. A significant challenge for UWB systems under the FCC ruling is the maximum power output. The FCC specifies a level of -41.3dBm per MHz, or about 0.007 mW per MHz. So, for a 528MHz wide signal, the maximum output power is still less than 10% of your basic Bluetooth headset, thus the extremely short range. And the regulatory climate in the rest of the world is not good and getting worse for WUSB, with the regulatory authorities outside the US only willing to allow use on the frequencies above about 6 or 7GHz, making the propagation issues even worse. Challenges from multipath and absorption by materials in the path create a system that may practically have only a meter or two maximum range. Right now, even though there are a number of companies that claim to have working, shipping, useful silicon, the real story is that it's a long way off from commercial utility, the price is very high, as is the power consumption. This is not a $5 chipset, nor should it be used in battery-starved devices. The Bluetooth SIG adopted the use of a form of the WiMedia WUSB radio for its upcoming Bluetooth 3.0 release, originally scheduled for the beginning of 2007, but now slipped out to (at least) October 2007, with real products not expected until 2009. The mandate from the BT SIG is that only UWB radios that operate in the >6GHz spectrum will be allowed, and the industry is years off from being able to meet that requirement. 802.11n is much closer to reality, its data rates are already comparable to the practical instances of WUSB, and it is not constrained by the exceedingly severe regulatory power limitations. It has a strong ability to trump anything that WUSB could eventually roll out, and crafted for very low power operation, be just as power efficient as WUSB.

They appear to be based on the G4, but have been redesigned:

"The chip contains three identical, customized IBM 64-bit multi-threaded PowerPC-based CPU cores, each implementing two simultaneous instruction threads and featuring clock speeds in excess of 3 GHz. The cores are enhanced with specialized function VMX acceleration for gaming applications, as well as with a high speed 128-bit vector unit. Additionally, the processor includes a 1 MB Shared L2 Cache with custom logic for high-speed data streaming for graphics and system applications, and provides an aggregate front side bus (FSB) bandwidth of 21.6 GB/sec, IBM added.

The chip is highly configurable and programmable, based on IBM's eFUSE technology. It contains 165 million transistors, and is fabricated in IBM's 90 nanometer Silicon on Insulator (SOI) technology to reduce heat and improve performance, according to IBM."

~nate

OSU says the near-invisible integrated circuit (IC) implements a five-stage ring oscillator, a function often used for testing and demonstrating new technologies. This is analogous to when software developers write programs that simply say "hello world," as an early step in testing and debugging new computer languages.

Hey, I know the girl on the very left in this pic! Just kinda weird... :)

caption from this picture in the article... http://www.deviceforge.com/files/misc/osu-transpar ent-electronics-team-big.jpg "This will only hurt a bit..."

Yes they are transparent! Check out this picture:
http://www.deviceforge.com/files/misc/osu-transpar ent-electronics-team-big.jpg

Precisely. I could EASILY get on of these into any data center into which I had physical access. I could nochelantly stick it into a USB port, a little bit of hocus-pocus with an autorun command, and I could easily accomplish the same thing without much effort at all. Hardly rocket surgery.

That is all well and good for corperate environments that need network access to programs from a server but seriously. This speed is 40 times faster then the connection I have at home for my internet. Unless you are doing things over your home network (Streaming video I suppose) there is no reason to upgrade.

The trouble is that theses companies will be pushing "N" routers like crazy when noone needs it. Unless it offers super Encryption of 802.11i then count me out.

Deja Bookshelf! Back in 1983, Silicon Valley startup Ampro Computers Inc. introduced a 'Bookshelf Computer' that consisted of a compact (7.3 high x 6.5 wide x 10.5 inches deep) cube-like box that could fit comfortably on a bookshelf along with book-sized expansion modules for additional disk drives and other peripheral functions. The expansion blocks interfaced to the main computer core via SCSI. The original Bookshelf Computer was based on a 4 MHz Z80 CPU running CP/M, while a second-generation model had an 8MHz 80186 CPU and ran IBM PC-DOS. Shortly after Ampro began shipping its Bookshelf Computers, Microsoft introduced a desktop software bundle it called the 'Microsoft Bookshelf.' Ampro subsequently informed Microsoft that its Microsoft Bookshelf product name infringed on Ampro's 'Bookshelf Computer' trademark, resulting in Microsoft purchasing the rights to the Bookshelf trademark from Ampro.

From the following article. http://www.deviceforge.com/articles/AT4903582708.h tml

"Full Legacy Support

With size issues, memory issues, platform compatibility issues, and more, desktop system vendors know that the BIOS train is rapidly running out of track. However, these vendors need a solution that not only resolves BIOS issues, but also allows legacy support for today's operating systems.

Intel's goal in creating the Framework for EFI was to help BIOS vendors support the speed, power, and innovations of today's system architectures. This solution offers full legacy support through the use of a compatibility support module, or CSM. For systems that do not yet have EFI, the support module takes the place of EFI so that the Framework can communicate with the traditional operating system.

The benefit to developers is that there is no real change to the work process. Even developers buying boards that include the Framework with EFI will not have to learn new or exotic tools. Instead, the same test suites, similar processes for handling BIOS, and so on, that have been used for existing systems can be used."

It sounds like EFI is fully backwards compatible with older OS'es but it will take a new BOOTLOADER. And, the best part, a bootloader that will boot Windows/Linux is included in OSX86!

-flipsoft

...but even so, to answer one of your criticisms, a colour version is indeed available (yes, linked to near the bottom of the original article!). Like OLEDs, it's going to be several years before these get cheap enough to consider using as an e-book (or whatever). I'm interested, also, whether this e-paper technology would scale up really large - e.g. could it eventually be used as a TV screen like they're eventually proposing for OLEDs?

I'd rather have one of these.

Doesn't anyone find it questionable marketing that the label says 1GB, but the USB stick says 512MB ? Take a close look at the big picture

Transmeta has a motherboard the size of a business card called Stratosphere.

Check it out here

They also have a desktop supercomputer, although I'll stick to my 800mhz fujitsu for now.

So shouldn't the 1750 be the 2100, or are they no longer trying to be even internally consistent?

Model numbering philosophy

AMD says its new model numbers are based on benchmarks developed by Synchromesh Computing. The scheme consists of the processor's family name (Geode NX or Geode GX) followed by its performance rating, followed by its power usage. Performance ratings reference performance relative to VIA's Centaur processors.

including photos and specs, are available in this news item at DeviceForge.

Let me summarize why Bluetooth is not dead, and wireless USB is not really a competitor for bluetooth:

Wireless USB Power requirements: 300 mw ("with a target of 100 over time")
Bluetooth power requirements: 100 mW, 2.5 mW, 1 mW (the last two are class 2 and 3, the variants widely used.)

Frankly, wireless USB sounds less interesting to me. Well, it's a threat to Wifi, from the sound of it. It's really, really fast and power hungry. It sounds primarily for unwiring our desk-bound, non-mobile computer peripherals from the computer. But then we will have to plug them all into the wall instead. So there are a few that had power anyway and now we've cut the number of cables from 2 to 1 - OK. But quite a few the only cable was USB (and that was providing power) anyway. It wouldn't be a viable solution for things like wireless mouses and keyboards, for instance. And I don't think I'd want that instead of bluetooth for the PDA/phone or PDA/computer link.

There are a lot of applications where very low power (1 mw!) is much more important than bandwidth.

Here's another article on DeviceForge which includes an architecture graphic, a list of sub-projects (and links to them), links to a presentation about Intel's Open Source Machine Learning software and a technical whitepaper, and a link to the SourceForge download site.

some may disagree, but i think its scary.

firstly tonight i read about Trusted Computing, and that Phoenix plan to put all sorts of weird and wonderful things into the BIOS (supposedly for our convienience and privacy, etc) Phoenix's BIOS Roadmap

and then i read about this DRM crap. It all seems to be tied together quite nicely, and results in a general lack of rights, ease of use, and privacy for the end user.

they are literally stripping away our choice, with this stuff. subtly making it more convienient to use proprietary things, (eg windows media player 9: you have to buy the mp3 encoding plugin, or your stuck with WMA... M$'s own audio thing, which does funnily enough have some crappy DRM "protection" in it)

hardware will become the same, im sure: maybe even Linux will refuse to run on certain hardware, or more realistically, the hardware will refuse to let linux run on it as its not trusted / secure enough, or doesnt have such and such a thing.. / cant do the encryption / decryption and all sorts of ..well, who knows.

i dont like it one bit...

anyways, im sorry for any typos, bad grammar, bad layout, presentation, etc.. i was never good at english, plus its my first /. post :D

Check this Wifi-car.