Slashdot Mirror

The markers they are using come from the open source ARToolKitPlus http://studierstube.icg.tu-graz.ac.at/handheld_ar/artoolkitplus.php for those interested... The demo marker is actually the one with ID = 0 ...

This is sad, just another example of how the wheels are coming off the cart while careening down another blind alley. I was at a trade show last month, and the visit to the Microsoft booth was surreal. The first kiosk was for Windows 7 and a smiling young man touting the virtues of this beta software. When I mentioned that I was having trouble running Vista on a 3.2GHz P4 with 4GB RAM, a 512MB ATI video card with DX10.1, and a terabyte HDD, he scoffed and said that nobody at Microsoft was running Vista, not even the developers. He gave me a DVD of beta 7 and told me that even as a beta, Windows 7 was "so much better than Vista." I accepted his disc (which expires on August 1), and went to the Windows Mobile (WM).

This kiosk had a good looking young man who was part of the product management group for WM 6.5 and very knowledgeable about the product. When I told him that I was a WM developer, he listened attentively as I explained my frustration in trying to program the WM6 smartphone camera to work. His smile faded as he explained that Microsoft had failed to thoroughly test the OEMs for WM5, WM6 and WM6.1. As a result, the DirectShow APIs for many phones were not fully/correctly implemented. He showed me a web page - http://studierstube.icg.tu-graz.ac.at/handheld_ar/camera_phones.php - that explained the problem phones. Then I asked, "will this be fixed in the coming 6.5 release?" He shook his head and replied, "no, not until WM7." I thanked him for his candor and moved onto Live Search.

At Live Search, a bright young man was touting the performance of their latest version and let me test it against Google, where it seemed to respond comparably. He talked about how his group was trying to get other parts of Microsoft to use their Live Search instead of their own, "an uphill battle." At that moment, another person walked up and asked a question, prompting him to pull out his iPhone. I reached out with my WM phone and joked, "wouldn't it be more politically correct to show this?" He responded, "oh, no. Most of my friends at work have iPhones. It's OK."

The problems documented by Daniel Wagner's web page (above) and unmentioned on microsoft.com or msdn.com cost us three months of development time. I should have suspected; mea culpa. Our application now runs on iPhone, and we are not looking back.

BTW, the Microsoft coffee table looks like a giant iPhone.

http://www.research.ibm.com/gsal/tcpa/tcpa_rebutta l.pdf

Palladium is a Microsoft led project to add "trusted" computing to Windows, through a combination of hardware and software. ... processor modifications to ... provide trusted path from the keyboard and trusted display.

http://www.microsoft.com/technet/archive/security/ news/ngscb.mspx

Secure path to and from the user. Secure channels allow data to move safely from the keyboard/mouse to nexus-aware applications, and for data to move from nexus-aware applications to a region of the screen.

To make NGSCB possible, both the software and the hardware will evolve. On the hardware side, the CPU, chipset, USB I/O and GPU hardware components will be redesigned, and a new component will be added, called the Security Support Component (SSC).

http://www.iaik.tu-graz.ac.at/teaching/03_advanced %20computer%20networks/ss2005/vo11/Palladium_LaGra nde.pdf

• Trusted USB-Input
• Old Version
• Trusted Channel from Keyboard/Mouse to Keyboard/Mouse Manager
• Need new Input-Devices which allow creation of Trusted Channels

New Version

• Update in Trusted USB-Input Specication
• Done by Intel and Mircosoft
• Protection Implementations in Chipset
• NO new Input-Devices needed

Protect Frame-Buffer from unauthorized access Transmission to and from Graphic-Buffer encrypted and protected from snooping

http://download.microsoft.com/download/1/8/f/18f8c ee2-0b64-41f2-893d-a6f2295b40c8/TW04055_WINHEC2004 .ppt

• Good news: Windows is designed to allow extensions and additions to this [the USB keyboard input] stack.
• Bad news: Things that extend or add to this stack are not always good news.
• Data manipulation
• Replay
• Substitution / Data Scaling
• Data leakgage

The article does a great job surveying some of the major players in the field. I think all of the cited researchers have received grants from the NIH Neural Prosthesis Program.

As mentioned in the article, BCI research is proceeding along invasive, intra-cortical lines as well as more data-processing intensive EEG-based approaches. The latter methods affix EEG leads on the scalp, record brain waves, and employ powerful computer methods to decipher the results. Noise is a problem, so researchers have embraced the more invasive approach of implanting chips directly into the brain. That's what Cyberkinetics and Neural Signals are doing.

The Lab of Brain-Computer Interfaces, Technical University of Graz, has an active group researching BCI, both through EEG and implanted electrodes. I'm surprised they don't get more press. There's also interesting work going on at Anderson's Caltech lab using the posterior parietal cortex, which might have some advantages. Check out the nice slide show on their research.

This is part of ongoing BCI research. The method involved is called Functional Electrical Stimulation, and FES control via a BCI is just one of the latest developments in this field. Take a look at this and this for more info on current developments.

More recently, a third option has become available: having fully implanted power system that can be recharged wirelessly, via em radiation of some kind (you can google for it). This is a big gain, because it allows devices that are more power hungry while still maintaining the benefits of not breaching the skin and not needing frequent operations. But it still requires people to remember and have access to the appropriate charging device consistantly. If for any reason some one forgets or can't recharge, the device may shut down, sometimes with fatal results. So having a way to remove one more step for powering these things should really help improve the quality of life for a lot of people today.

Of course, personally I find this to be a very exciting development for future things as well. When we get to the point of having more optional implants, for things like boosting hearing or vision, a way to power them will be necessary, and if the power requirements are low, then this system would be perfect. Ultimately, widespread adoption of anything, from an OS to a vehicle, is all about making it as easy and intuitive for end users as possible. There is a lot of interesting stuff going on for advanced things like brain-computer interfaces, and people who are interested should look around, as the state of the art has advanced a great deal in the past 5 years. Here are a few links for the curious, and much more can be found with google, of course:

Graz University of Technology

Standford/DVA Neural Interface Project

Beyond the Big Barrier(lighter, intro type stuff)
News Group:
sci.med.psychobiology

So, what kind of "innovations" has been created by Microsoft? Maybe Clippy. But that's it, and we all know how helpful that is...

Never, ever, forget the badness that was M$ Bob. According to reports, Bill's future wife was the project manager. I have no idea what that says about either of them. The illusion of security through not even trying to be secure.

The designers of other languages like Perl, C++, Java, etc, didn't have a good enough understanding or appreciation of Lisp, or those languages wouldn't be the disasters they are today. In fact, many of the "pop" languages have displayed an almost homophobic contempt for Lisp.

Vanilla Ice surely knows about Miles Davis, just like Stroustrup known about Lisp, but he certainly doesn't draw any signifigant influences in a way that shows up in the quality of his own work, in the way that Python draws many infliences from Lisp.

Please explain how anyone who actually understood the true power of Lisp macros, could have ended up designing something as awful and disgusting as the C++ template system? (And he still left in all the fractal syntactic interactions and subtle unintended toxic side-effects of the C preprocessor.)

Since you say the burden is on me to prove how continuations are not am important part of Lisp, I refer you to the fact that continuations are not part of Common Lisp. Kent Pitman (who helped me learn lisp) writes:

"I, for one, am glad not to see continuations in CL. If someone proposed adding them, I would argue strongly against the move. I feel about them the same way as others feel about GO. They encourage unstructured programming and I think they're a semantic mess."

Scheme is very different from Common Lisp, in many signifigant ways, yet nobody would argue that it's not influenced by Lisp. Python is influenced by other languages than Lisp, but Lisp is certainly one of its main influences, and it's obvious that the designers understood Lisp well, and didn't hold it in contempt like the designers of the Perl or C++ template system.

I think one of Lisp's greatest problems is that its name has homosexual connotations, and that scares off all the homophobic closet queens, who don't even realize why they're supernaturally subconsciously afraid of the language. I know a lot of gay Lisp programmers, which suggests that you have to come to terms with your sexuality before taking on a language with a name like that.

-Don

AutoLISP is the scripting language for AutoCAD by Autodesk, a very crippled dynamic Lisp-1, based on very early xlisp sources (v1.0), posted by David Betz on usenet (alt.sources), and without proper copyright laws that time used by Autodesk as their free scripting language. There's an old Byte article about that.

To keep the language simple, some typical features such as macros, vectors, structs, destructive operations and objects were left out. Memory was low then in the late 80'ies. Since then the language itself was not improved at all, to keep things simple and to keep backwards portability....

By "dynamic" I assume it means dynamically scoped. Lexical scoping is an important feature, introduced by Scheme in the 1970s, that today is the unquestioned best way to scope variables. Perl and Python were both dynamically scoped when first released, but switched to lexical scoping later.

Macros let you customize the language to specific application domains, resulting in very readable code. They are a feature that is difficult to duplicate in languages without lisp-like syntax. C preprocessor macros are nothing by comparison.

Autolisp missed out on a lot by not having macros. Also objects. The Common Lisp Object System (CLOS) is one of CL's most vaunted features.

I'm not saying that a LISP dialect would have been better for this particular project. It certainly fails on the "Similarity to C" qualification. Just don't take your experience with autolisp as indicative of what LISP programming is like.

DS1's AI ran on more than "just" Suns. The probe itself runs VxWorks on a rad-hardened RS6000 (PowerPC) at the awesome speed of 25 MHz (see here).

Development of the planning/scheduling was done on Suns and Power Macs, using two different vendors' Common Lisp implementations (see here for a message from one of the implementors). During development, NASA management decided there were too many programming languages flying in DS1, so they decided to drop one of C, C++, or Lisp. C++ lost, but is being wedged back in for political reasons.

The planner was only given 10% of the CPU, which meant DS1 was doing real-world AI at 2 MHz (!).

Remote Agent, written in Common Lisp, wins NASA's Software of the Year 1999. Dissatisfied, NASA management decides the entire project should be ported to C++ !!?

Chuck Fry, one of Remote Agent's developers, says:

"The bad news: the Remote Agent modules are all being ported to C++ anyway, as C++ is seen as more "saleable" to risk-averse flight software management. This despite the fact C++ was bumped from Deep Space One because the compiler technology wasn't mature, which the Remote Agent planner/scheduler team learned to its dismay."

Ack! Looks like you are right, the Mac version is way behind and no longer linked. This post explains it, sort of. Looks like Dave ran into some problems doing the port quite awhile back, since he relies on Novell libraries, and Novell quit supporting Macs. From the date of the post one might deduce that the problems proved insurmountable, or at least more trouble than he thought it was worth.

Not being a Mac user I remained blissfully ignorant of this till now. Sorry :(

Now THAT sounds like an entertaining project. And, since it's protected under the GPL, the authors would release those extensions...