Slashdot Mirror

Zone Logic is a nice combination of state charts and autononmous agent design. It was used in the 80s and early 90s for industrial automation, and was particularly good at recovery from hardware failures.

Real-time systems depend on sensors to report the state of the outside world. These sensors often fail, putting the system in an illogical state. Zone Logic offered a nice structured way to deal with this.

R. Roberts. Zone Logic: A Unique Method of Practical Artificial Intelligence. Radnor, PA, Compute! Books, 1989. citeseer

Allan Snavely... hmmm, sounds fishy to me - does he have good credentials?

Ok, hompage, Paper on scheduling threads on one processor, Write-up on speech about aforementioned paper... he sounds legit, but that name? Snavely is like sly (S nave ly) and knave (minus the "k") mushed together... I don't know if we can trust his recommendation.

For most of theses transformations to be correct, the compiler has to prove there is only one pointer to the object -- in the whole program. Whole-program analyses are expensive, and so are point-to analysis. And there is just not that kind of time to spare in a JIT, where every second spent analysing the program is time spent not executing it.

The optimization the book proposes are all hit-or-miss adventures. Even for a programmer with intimate knowledge of the code, it is sometime difficult to predict if a change will help or imper performance. The compiler has even less chance to do so correctly -- and nobody like a compiler which slows down their code trying to optimize it.

Once you find a paper that is close to what you want there are ways to surf to similar/related papers.

Once you find a paper that is close to what you want there are ways to surf to similar/related papers.

If somebody intent on breaking through the smart card's security has access to the smart card, then sooner or later the security WILL be broken.

Some pointers:

• Smart card FAQ
• Design principles for tamper resistant smart card processors
• Low cost attacks on tamper resistant smart card devices

If somebody intent on breaking through the smart card's security has access to the smart card, then sooner or later the security WILL be broken.

Some pointers:

• Smart card FAQ
• Design principles for tamper resistant smart card processors
• Low cost attacks on tamper resistant smart card devices

And the most scary part of this plan is the "cooperative" nature of the suggested radios. Creating a worldwide infrastructure without considering the security holes is negligence. How nice, the signal you receive will contain a URL to tell you how to download new software for your radio. And if that signal is from an illicit source telling your radio to start relaying spam back onto the net?

Reed's analysis, badly presented in Salon, deals with networks of wireless nodes that not only use frequency diversity (e.g. spread spectrum), but also use multiple antennas for spatial diversity (e.g. phase arrayed antennas) and the nodes cooperate not only for relaying (e.g. mesh network) but also for detecting and eliminating interference.

All of these elements increase the efficiency of radio spectrum use.

Optimal Operation of Wireless Networks

Combined Space Time Diversity and Interference Cancellation for MIMO Networks

Information Theory at the Extremes

Linear Multiuser Receivers: Effective Interference, Effective Bandwidth and User Capacity

Abstract: Multiuser receivers improve the performance of spread-spectrum and antenna-array systems by exploiting the structure of the multiaccess interference when demodulating the signals of a user.

The original article that introduced it can be found here (download links for the article in several formats are at the top). Google also has a nice collection of links.

However, RC4 has had some setbacks recently.

A paper on attacking RC4
A more theoretical attack paper
A paper describing an attack that requires some guessing and probability theory

I have a few problems with the RC4 algorithm, only one of which I'll talk about. It's not implemented poorly in one or two protocols, but several. If it's that hard for engineers to implement properly, then my brain simply thinks "Don't use that protocol! It's not worth it!" Perhaps, it's an okay protocol. Perhaps, there are just too many engineers that don't know how to implement it properly.

Thank you for your time,

Quadgoatboy

P.S. As a side note, I don't know about you, but I just don't trust a protocol that padds itself 64 times on a 32 bit key. That just kind of... creeps me out :D. Yeah, yeah, I'm paranoid, but isn't anyone who uses encryption.

This site is for searching CS publications based on citations and other criteria. There are numerous papers on quantum computing many of which provide an excellent introduction. Ofcourse the webpage of faculty and their lecture notes ( here and here ) provide an excellent introduction. I would recomment going through the lecture slides before attacking a few of the more readable and fundamental papers.

Check out CiteSeer's links: CiteSeer. This problem has been known for quite a while. Check out the MOAS problems with BGP, though, if this isn't enough to worry you.

Jouster

Well that would, of course, depend on how much bandwith is running to said server now wouldn't it? =)

Found some nice info (good old google) on said Supercomputer though since the sites linked article didn't have much.

A Time Article on The Earth Simulator

Top 500 page on Earth Simulator

NEC page on the Earth Simulator

Google Translated Powerpoint presentation on the Earth Simulator

A snippet(s) of info:
"Based on the NEC SX architecture, 640 nodes, each node with 8 vector processors (8 Gflop/s peak per processor), 2 ns cycle time, 16GB shared memory. Total of 5120 total processors, 40 TFlop/s peak, and 10 TB memory. "

"Earth Simulator's processors are one-chip LSIs fabricated with 0.15 micron CMOS process and copper wiring. Highly optimized software and high-speed networks that pump massive amounts of data through 7.8TB/s bandwidth connecting the 640 processing nodes are key to the amazing efficiency of Earth Simulator."

. . .

OK, I use these guys. I'm happy with them but obviously biased.

Check their NG archive : here and see what you can find.

In summary, not only will AAISP happily set you up with a v6 alloc (being a RIR) but they and their users are getting quite clued up on IPv6. AAISP's bias is towards Linux, so you may be happy in their company.

Their v6 is tunneled to your site, because of limitations in the BT backhaul (a ATM cloud owned by our local monopoly) but this may change with SDSL + local exchange unbundling within the next 12 months.

Of course, one's often quite limited in delivering v6 direct to hosts on your own net' party due to IOS inconsistencies (and I'm not certain Cisco or anyone else has complete v6 across their products) - but these NEC v4+v6 router - switches look very attractive IMO.

Just some idle random thoughts, but I'd rate Andrews and Arnold as a "IPv6 Friendly" ISP. There are many other was to get yourself an allocation and tunnelling to peers or the 6bone however.

it does not scale across multiple processors. You get the same performance on one versus eight CPUs in the benchmarks I've seen

And these would be which benchmarks, exactly?

SPECjbb 2002 Q1 and SPECjbb 2003 Q3-- look at the xSeries 360 and 370 results, where it scales quite nicely as you double the number of processors.

More likely it's just some sychronisation primitive(s) that they need to tune or remove

Holy shit, let's all rush out and tell David Bacon that his locks suck :)

Well, I may have overestimated the potential for convergence :-)

I think you misunderstood. I used to be a "Java fan" and am responsible for its adoption by several companies.

I see.

But Sun has demonstrated bad faith and incompetence when it comes to Java over the last half dozen years: not only has Sun patented key aspects of Java, they have also pulled out of several standardization efforts, and they have failed to deliver essential technologies and enhancements that they promised.

But it seems that Dotnet is the lesser of two evils after all, and people following your earlier advice to embrace C Sharp are now the ones who have cause to worry.

Sun, in contrast, have continuously opened up the specs, tools and standardization processes over the last year, as OSS developers are happy to acknowledge. Yes, they still retain patents, so in deciding who is the lesser evil one has to compare MS's record with Sun's. Two points in Sun's favour are that they've allowed the Kaffe and JBoss people access to the JVM and J2EE test suites (previously licensed) and that IBM, BEA and a bunch of other implementors are apparently happy to continue investing in Java (TM) and working with Sun.

I find it rather ironic that Swing, the GUI that Java-baiters love to hate, is now elevated to a must-have API. It seems that when it is present people demand an alternative; when it isn't, AWT and SWT are somehow no longer acceptable.

There is still some hope for C#: the Mono project is actually increasingly relying on non-.NET APIs

So you're saying that the less hope there is for Mono cloning Dotnet, the more hope there is for C Sharp? I trust that those who are still following your advice can keep this principle clearly in mind.

And so to the collaborative development of a new programming language, which appears to have got off to a rather rocky start:

1. Java-style byte codes are an awful representation for manipulating programs.

Quite right, and one reason why we're not talking about Java, but about an alternative to Java.

2. Programs can manipulate persistent data directly rather than mapping it to and from storage systems.

Well, not in Java, and not in anything with a Java runtime.

You're really not paying attention are you? Java isn't relevant here, something like SQL might be.

3. Commercial workflow has nothing to do with operating system processes or threads.

From whose perspective? Not from mine, if I'm programming logic using BEA WebLogic Workshop. This allows me to express my program as a sequence of steps that can be distributed over a number of machines, persisted, paused and resumed, managed and interrupted.

The goal of high-level programming systems is to abstract and automate - abstract away from system details such as workflow to thread mapping and automate processes such as memory management.

4. I have no idea whose "original intent" for LISP that is supposed to have been.

Indeed? My view is that the history of LISP is highly relevant to modern language design. As several people have observed, all language developments tend to converge towards something like CLOS. For your reference, the original intent that LISP be an intermediate language is described here.

I think multi-language support is vastly overrated. [...] I do think a platform should support mixing high-performance statically typed code and convenient dynamically typed code, but for that, you only need two languages (java/bsh, C/Tcl, C++/Python, etc.).

I take it that you don't use SQL or XML?

I agree that Dotnet-style "multilanguage" support is overrated, but that's a limitation of the CLR, not the principle.

Regarding static and dynamic typing, it is a mistake to assume that the use of both models must or should imply separate languages. This paper, and implementations such as Strongtalk (on Smalltalk) and typed modules (on Scheme) show the benefits of unifying the approaches, which are after all nothing more than (partial) program verification techniques.

4. I don't want security features in my day-to-day language: they are complex and costly.

Obviously your requirements are somewhat different from typical corporate IT systems. That's perfectly OK, we will keep this in mind when evaluating any future recommendations.

Java is not a particularly well-engineered platform because many of its tradeoffs were driven by one environment (platform-independent, untrusted client software) and make no sense for a general-purpose language.

On the contrary, a VM must be able to deal with untrusted code as well as trusted, otherwise it wouldn't be general-purpose.

The Java VM supports different levels of trust, allowing code from different sources to coexist. What's more, it can do so efficiently, since the JIT allows things like access checks that are always true to be optimized out - something impossible in the purely compiled languages you seem to be advocating.

And C# has copied most of those bad tradeoffs.

Well, it missed the opportunity to innovate, we can agree there.

Perhaps it's good that both Java and C# are removing themselves from the space of open, free languages: it might be best to start over with a simpler, better engineered system anyway.

Absolutely.

That's a priceless comment. Do tell how one can have cause and effect without the notion of time.

Time is a property of space-time geometry. Causality is not. Go get some clue.

The NoW (Network of Workstations) approach has been on ongoing trend over the last few years as the throughput achieved by an N distinct processors connected by a high speed network is nearly as good (and sometimes better) than an N processor mainframe. All this comes at a cost that is much less than that of a mainframe. In Google's case, it is the volume that is the problem, and not necessarily the complexity of the tasks presented. Thus, Google (and many other companies) can string together a whole bunch of individual servers (each with their own memory and disk space so there is no memory contention - another advantage over the mainframe approach) quite (relatively) cheaply and get the job done by load balancing across the available servers. Replacement and upgrades - yes, eventually to the 64 chips - can be done iteratively so as to not impact service, etc. Lots of advantages...

Here is a link to a seminal paper on the issue if you are interested:

http://citeseer.nj.nec.com/anderson94case.html

Aren't there enough papers already on IPv6? Especially on purpose #1 (i.e. increasing the internet experience).

For #2 (i.e. promoting widespread), it's highly debatable, IMHO...

It's not just dynamically loading a class.

It's dynamically loading a class with a number of conditions: type-safety, lazy runtime checks and loading, class-loading customization, isolated contexts/namespaces with their own classes and security policies, etc.

As far as I understand, the issue is that what Java provides requires a lot from the run-time system: type-safety checks, security checks, etc. which is provided by the VM.

I don't know if this could be provided by compiled code, but my guess is that if it could, it would duplicate most of the VM into a runtime system for the running program. So each executable would bring a hefty and redundant RE along with it.

The difference, it seems to me, is that Java assumes the code to be dynamically loaded to be unsafe and to come from an unknown source (network), and loaded through an unknown protocol (custom class loading), under unknown security constraints, all of which can be defined at runtime. So the VM has to be on top of everything.

I think other compiled programming languages are more trusting with their dynamically loaded modules, so the runtime environment code is probably less cumbersome.

I could be wrong, but I think the other languages that offer "safe" dynamic loading also have VMs.

I just found This paper, which could be helpful.

I don't think Matsuis paper is available for free online. If you have access to a good (university) library you can find "Linear cryptanalysis method for DES cipher" by Matsui in Advances in Cryptology, Eurocrypt '93. The results were improved in "The first experimental cryptanalysis of the Data Encryption Standard", published in Crypto '94 by Springer Verlag in the "Lecture Notes in Computer Science" series.

Biham wrote "On Matsuis linear cryptanalysis", available at http://citeseer.nj.nec.com/biham94matsuis.html which contains a nice overview.

Terry Ritter has a literature survey at http://www.ciphersbyritter.com/RES/LINANA.HTM

http://citeseer.nj.nec.com/35491.html is the paper by Biham & Shamir on the differential attack on full DES.

I don't think you could narrow the keyspace in this way - you really need a right pair and then you find the key.

The plaintext is actually 8 bytes - it's the key that is 64 bits with 8 parity bits giving 56 bits effective.

I think Gutenberg is very much there... Have you ever looked at the amount of material in Gutenberg's archives? When it comes to books and material written in english, that is in the public domain, I have to say, that Gutenberg offers almost everything of interest already.

The 'vision' that the author of the Wired article had was somewhat different: To be able to access all texts electronically. Something that everybody who had to hunt down old magazine articles has dreamt of (I still have nightmares from that one dark and dusty university library cellar, *shudder*). While Gutenberg is a great project, to come closer to full availability of all texts via electronic media, there will have to be initiative from governmental organizations as well as commercial entities. Obviously, not all texts will be available for free. But even a somewhat unified way of searching and finding these texts will be huge task.

There is CiteSeer for articles on computer science, there is IEEExplore if you happen to be looking for something from IEEE. But you have to know these places. Even with better search engines like Google it's still quite a task to get your hands on a text, even if you have some time to do the search and are willing to spend money.

A large database of text references (maybe including abstracts) would also be nice to just see what's available while you are still doing research.

The reason the Gutenberg project isn't hugely succesful is not the lack of text. Part of it might be the lack of formatting. Nobody want's to read 600 pages of a classic work on a computer screen in ASCII.

GutenMark does that (almost) automatically. Uses LaTeX.

I guess you could argue that it was unnecessary--but I was the one laughing when everyone else thought "the internet was down" when they couldn't get to Yahoo Games, Hotmail, and Google.

Of course, it might be noteworthy to mention I find far more "relevant" uses of my Internet connection. :)

The 3000 gate figure is just the nand gate equivalent count, in reality it only has 128 moderately configurable cells. By comparison the FPGAs that people use for anything slightly serious start at about 600 cells (~40k gates, the cells in FPGAs are more complex). In the stuff I do (image processing mostly) the smallest FPGA I use is a xcv1000e with 27,648 cells (~1.5M gates), and this is fairly out of date.

OTOH you really don't want an FPGA in a handheld, as they are very inefficient in terms of power/heat. They get way too hot to touch running in my computer (admittedly without active cooling, but that would be difficult to achieve in a handheld anyway), and suck a huge amount of power (some of our machines with weaker 300W PSUs can't provide enough current to boot with an FPGA card in). Admittedly the parts I use are more heavyweight than you'd think of putting in a mobile device, but even the smaller ones need alot of power.

Annapolis (a slightly odd FPGA platform vendor) do a PCMCIA card called the Wildcard which apparently works in an iPaq (plus adaptor), for example used in this project for wireless encryption project. But they cost $999 :)

Abstract:

Memory Management Units (MMUs) are traditionally used by operating systems to implement disk-paged virtual memory. Some operating systems allow user programs to specify the protection level (inaccessible, readonly, read-write) of pages, and allow user programs to handle protection violations, but these mechanisms are not always robust, efficient, or well-matched to the needs of applications. We survey several user-level algorithms that make use of page-protection techniques, and analyze their common characteristics, in an attempt to answer the question, "What virtual-memory primitives should the operating system provide to user processes, and how well do today's operating systems provide them?"

Please see the following paper for motivation and examples:

http://citeseer.nj.nec.com/536115.html

(See the links in the top right to download/view.)

Amir

I used DUEL about 5 years ago. It's an extension to gdb and similar in design to Acid. I am not sure if its features have been integrated into GDB now.

Interlisp in all its incarnations had a module called DWIM (Do What I Mean) which caught typos and thinkos in your input and offered to correct them for you. It was at its smartest fixing spelling errors and inserting parentheses (it would guess '(foo)' when you typed 9foo0, for example).

DWIM could be turned off, set so it asked you before changing anything, or set to automatically fix your inputs. That last setting would be completely intolerable without UNDO, which Interlisp also had, so even if DWIM guessed wrong, corrected your input to the moral equivalent of 'rm *', and ran it, you could still recover with UNDO and try again.

The oldest Interlisp reference I can find is the 1978 Interlisp reference manual. Lisp was on Altos as early as 1973, hacked up by L. Peter Deutsch. For an academic take on UNDO, see this paper from ACM TOPLAS (courtesy of Citeseer).

Vector subtraction implemented neurally: A neurocomputational model of some sequential cognitive and conscious processes

Artificial synapses copy brain dynamics

Computing and Learning with Dynamic Synapses (1999)

Computing at the Tissue / Organ Level

From neurobiology to silicon

Principle of Neuroinformatics and Neuroinformation Coding


In short. The synapse is a computer in it's own right.

Read Handbook of Applied Cryptography (pretty much the definitive source for number theory as used in cryptography) - Algorithm 4.6.2 is entitled "Maurer's algorithm for generating provable prime's" and disagrees with your point "The only way to generate a real prime is factor it".

Oh, and how do you "factor a real prime"? ;))))

Yes, Knuth vol 2 is a good reference - but it is lacking in this respect. It doesn't even mention Maurer's algorithm.

I would suggest reading up on Maurer's paper and then writing to Mr Knuth. :).

"number of primes is in the realm of 3.7e151" - yep, that's for 512-bit primes.

Citeseer is the best free computer science digital library. Every computer scientist should have that site bookmarked!

It takes about 15 minutes to crack 64-bit WEP. A day to crack 128-bit WEP. I think that 256-bit WEP IIRC would in theory take about a month of non-stop monitoring.

As for your figures, they are incorrect. Fluhrer, Mantin and Shamir have shown it is possible to "recover an arbitrarily long key in a negligible amount of time which grows only linearly with its size, both for 24 and 128 bit IV modifiers".

Thus, doubling the size of the key will only double the time needed to recover the secret key! Stubblefield, Ioannidis and Rubin have shown that it takes as little as 1,000,000 packets to recover a "128-bit" WEP key. That's less than a day on a lightly loaded WLAN to recover a hypothetical "256-bit" key...

The problem here is that for DRM to work, it has to be under control by one company or consortium - like the DVD CCA

Hmm. You may be right about that. I'm not entirely convinced that there is no way to have an open-source implementation of DRM. I don't really know anything about cryptology, but I don't see why in principle DRM couldn't rely on PGP type authentication, i.e. that the implementation could be made source-code available, and the "rights" could be encoded in the sender's certificate without a third party authentication repository required. However, this PDF paper by Intertrust seems to conclude, in somewhat self-interested fashion, that a trusted 3rd party is necessary. On the other hand, in the real world we see the ongoing development of OpenPIMP, excuse me, OpenIPMP which seems to contradict your assertion. An interesting point, nonetheless, and I think you deserved to be modded up.

I think DRM/trusted computing is inherently evil.

I think you and I and about 99% of the Slashdot community are in agreement on this. The remaining 1% consisting entirely of Twirlip of the Mists and his varies friends and fans.

...on why, perhaps, folks get slashdotted:

-rw-rw-r-- 1 jdf jdf 57197 Dec 17 10:12 original.html
-rw-rw-r-- 1 jdf jdf 10425 Dec 17 10:28 text.html

Yes, that's 47 kilobytes of javacript, headers, footers, CSS, and associated crap for about 10K of real info. Sure, there's also pictures and such, but this is basically 4x the original in overhead alone. Anyhow, here's the text from the site:

Pick a toy--any toy--for holiday gift-giving.

By Joan Lynch, Maura Butler, and Matt Miller -- EDN, 12/12/2002

Holy spokes, Bikeman!
Attach Hokey Spokes to your bike, and you'll have not only the coolest ride in the hood, but also the safest. Inventor and MIT graduate Dave Hoch wanted to take the terror out of night bike riding and designed the transparent "blades" to wirelessly communicate patterns or words with each other. You can put up to six blades on each wheel; a computer on each strip controls 16 LEDs in a variety of colors. Each blade costs $29.95 at www.hokeyspokes.com. Illumination Design Works, 1-617-924-4014.

Keyboard portability on a roll
Road warriors who prefer a full-sized keyboard with their laptops or PDAs will appreciate the Manhattan True-Touch Keyboard. When you're on the move, you can roll up the flexible, plastic unit and tuck it away. The keyboard features a QWERTY design and 104 raised keys that provide goodand quiettactile feedback, but using them will take some practice. Choose from PS/2 or USB connectivity. Units are dustproof and moistureproof and offer a lifetime of 15,000,000 keystrokes. Keyboards are available for around $27 from ThinkGeek, www.thinkgeek.com.

The "PC" PC
In an environmentally friendly move, NEC designed its PowerMate eco PC with 100% recyclable plastic and a 15-in. LCD flat-panel screen that contains no boron. The motherboard is made with lead-free solder, and the computer has no fan, which reduces the amount of dust dispersion for users with respiratory problems. The eco contains a 900-MHz Crusoe processor and a 20-Gbyte hard drive. And of course, it's Energy Star-compliant. The eco starts at $1599. NEC Corp, 1-800-338-9549, www.nec.com.

Free space
Combining a DVD player with a hard-disk-based PVR (personal video recorder), the Scenium Digital Media Recorder (DRS7000N) could be just the thing for overcrowded home-entertainment centers. The $599 unit features progressive-scan output for the clearest pictures and plays a variety of discs, including MP3 CDs and DVD+Rs/RWs. The PVR section handles the usual trick features, such as pausing live shows, and offers more than 30 hours of recording on its 40-Gbyte disk. Best of all, unlike PVRs such as TiVo, there's no fee for the EPG (electronic-program-guide) service. RCA/Thomson, 1-317-587-3000, www.rca.com.

The power of pictures
CyberLink's PowerProducer lets you produce your home movies and photos on DVD or CD for sharing with friends and family. If you're completely inexperienced, a step-by-step wizard-style program eases the way. It takes just three steps to import photo or video files from video-capture devices; you can then add special effects; adjust color; and trim, merge, and split video. The "QuickBurn" feature captures video from DV camcorders directly into DVDs in one click. Prices begin at $49.95. CyberLink USA, www.gocyberlink.com.

Mouse exterminator
Rid your desktop of mice with the iGesture Pad. The touchpad, which is just a bit smaller than a mouse pad, lets you use finger gestures to complete the same operations you would with a mouse. The $179 pad recognizes hundreds of hand gestures, including those that let you point, click, drag, and scroll. Twist your hand to open a file; spread your fingers to zoom. The USB-connected device is thin and small enough to travel and accepts commands from the right or left hand. Hmmm, wonder how it handles gestures you might make when your computer isn't behaving the way you'd like it to? FingerWorks, www.fingerworks.com.

Carry all
The multifunction Duex is an MP3/WMA player, voice recorder, and data-storage device in one portable package. Take off the bottom of the device to find a USB plug for attaching to the appropriate port on a computer. You can drag and drop music, image, video, and data files from the PC to the Duex and vice versa. The device features 128 Mbytes of memory for two hours of MP3 playback, four hours of WMA playback, and eight hours of voice playback; a backlit LCD shows song information in numerous languages. One AAA battery gives you more than 12 hours of music play. With a headphone/neckstrap, USB cable, software CD, manual, and one battery, the Duex mp302 costs $179. Innogear,
www.innogear.com.

Click it, stow it
The ultrasleek Pocket Digital is a handy, stylish way to capture favorite moments. The credit-card-sized digital camera captures and stores 52 high-resolution images at 1.3 megapixels. The lithium-polymer battery endures for hundreds of pictures and recharges through the USB connection. Image downloading is a snap. $129.95. Logitech, www.logitech.com.

The end of e-mail angst?
Neo (Nelson Email Organizer) might be just the ticket for reducing the stress associated with your daily barrage of e-mail. This add-on for Microsoft Outlook lets you work alongside or outside Outlook. Neo automatically sorts and prioritizes messages into intuitive folders. Searches are quick, and complicated filters are not necessary. The program deals with spam and bulk mailings and manages groups of messages by the type of attachments they contain (for example, all Excel files or Word documents together). Neo costs $39.95 for one copy; site licenses are also available. Caelo Software Inc, 1-250-354-5580, www.caelo.com.

Power protection
As more and more electronic equipment finds its way into more and more households, UPS devices could become commodity items for the average home owner. Six UPS models from Energizer will help protect home offices, home theaters, and PCs from power surges, spikes, brownouts, and outages. They automatically save open files, safely power down systems, and protect hardware when threatening power irregularities arise. Features include visual and audio warning indicators and USB connectivity. Prices range from $59.99 to $279.99. Eveready Battery Co, www.energizerups.com.

Surreptitious snapper
Fitting into the palm of your hand, the Mini Pen Cam 1.3 offers a still-image resolution of 1248960 pixels. Using its 16 Mbytes of flash memory, the device can store as many as 50 full-resolution pictures or 160 snapshots snapped at VGA resolution (640480). The $79.99 gadget also gives you the ability to shoot AVI-format movies, although their quality tops out at 624480 pixels and a choppy 10 frames/sec. In addition, the device, which runs on two AAA batteries, connects to a PC via USB to function as a Webcam. Aiptek, 1-949-585-9600, www.aiptek.com.

In touch, on the go
BlackBerry now lets you make and receive phone calls. The palm-sized BlackBerry 5810 wireless handheld device operates on GSM/GPRS networks to allow communication via wireless e-mail, SMS (short messaging service), or integrated GSM phone services, including call waiting, call answering, conference calling, and call forwarding. It comes with an earpiece and microphone for hands-free operation. You can even click on a telephone number inside an email message to place a call. The $499 device incorporates Java 2 Micro Edition. Research in Motion, www.blackberry.net.

Dude, you're gettin' a PDA
Like the PC before it, the PocketPC platform continues to attract more and more manufacturers. PC powerhouse Dell recently joined the fray with an aggressively priced model called the Axim X5. The device features a 3.5-in. transflective TFT with 240320-pixel resolution, as well as both CompactFlash and Secure Digital expansion slots. Two variants are available. The $299 model has a 400-MHz XScale processor, 64 Mbytes of SDRAM, and 48 Mbytes of flash ROM. A $199 configuration steps down to a 300-MHz processor and 32 Mbytes of each type of memory. Dell Computer, 1-800-999-3355, www.dell.com.

Risk averse
If you carry your office in a pocket-sized device, you're asking for trouble. What if you leave it on a plane or in a cab? The iPAQ h5450 tries to reduce the danger with integrated biometric security. A thermal fingerprint reader means only you can access your priceless data, and you can even add a layer of safety by combining the fingerprint sensor with a password. The $699 pocket-sized PC also includes a 400-MHz XScale processor, built-in IEEE 802.11b and Bluetooth capabilities, and an SDIO (Secure Digital Input Output) expansion slot. Hewlett-Packard, 1-650-857-1501, www.hp.com.

Bragging rights
The Zaurus SL-5600 gives the gadget geek ample reason to crow. It not only features a cool physical design with a hideaway keyboard, but also runs a version of Linux. The PDA boasts a 400-MHz Intel XScale processor, CompactFlash and Secure Digital expansion slots, 64 Mbytes of flash, 32 Mbytes of SDRAM, and a juicy 1700-mAhr battery. The device is slated to appear early next year, and pricing hasn't been announced yet, but its predecessor, the SL-5500, currently goes for $375. Sharp Electronics, 1-201-529-8200, www.sharpelectronics.com.

Stop graffiti
Thumb-pecked keyboards have become so popular on handheld devices that even Palm, the originator of the Graffiti handwriting-input system, now offers them. The $549 Tungsten W, slated for the first quarter of next year, features a tri-band GSM/GPRS radio and supports phone calls (via a headset), Web browsing, e-mail, and messaging. Powered by a 33-MHz Motorola Dragon-ball VZ processor, the PDA includes a Secure Digital expansion slot, Bluetooth, and a 320320-pixel color display. Palm, 1-408-503-7000, www.palm.com.

Is this your study?

It doesn't support your point. In fact the abstract says "The rate of acquisition of new links is probably proportional to the number of links the site already has, because the more links a site has, the more visible it becomes and the more new links it will get." Yes, new sites have the ability to quickly get a lot of links, but without the benefit of being mentioned by the existing media, it's hard to get the visibility which is required to get the links.

In the field of computer science, this problem is probably less pronounced. Several resources make it very easy to read papers, so access is not much of an issue. These resources include CiteSeer and the ACM Digital Library, which contain vast databases of computer science publications in electronic form.

If people are going to use this archive to automatically induce rules for recognising junk mail (e.g. using naive bayes or ripper), then they will also need at least as many examples of legitimate mail.
Of course it could be useful for evaluating classifiers built using smaller corpora.

If people are going to use this archive to automatically induce rules for recognising junk mail (e.g. using naive bayes or ripper), then they will also need at least as many examples of legitimate mail.
Of course it could be useful for evaluating classifiers built using smaller corpora.

> Someone here did a project last year to "derive" a new symphony by a composer. The idea was to analyse various pieces written by the chosen composer, find the common themes, and then use them to produce new pieces which would have the same "feel" as the originals.

Here is a link to a paper a guy(?) wrote about using neural networks to create fake Bartok melodies. Follow the links for more along the same lines.

Of course, Bartok always sounded like sequences of random keystrokes to my Philistine tastes, so I can't judge how well the imitation worked.

In this article they describe a quite nice system that help this spam filtering and also helps anonymizing registration in websites and does password management, all using strong cryptografy. Quite nice idea, though I haven't used the actual implementation.

Consistent yet Anonymous Web Access with LPWA

Citeseer is one of the best free online Computer Science digital libraries. If you are ever doing research in CS, check out Citeseer first!

"Articles freely available online are more highly cited. For greater impact and faster scientific progress, authors and publishers should aim to make research easy to access"

"Articles freely available online are more highly cited. For greater impact and faster scientific progress, authors and publishers should aim to make research easy to access"

That Joel! Always coming up with catchy phrases for concepts that his betters have already published papers on. See David Keppel's 1993 paper for a more thorough explanation, complete with lack of baby-talk.

This approach is more commonly called "Naive Bayes" classification in the field of machine learning. It is naive because it considers each word to be a feature (dimension), but it also considers each word in an email to be conditionally independent of all other words in the document (which is not true, but really useful in practice).

The author of the web page on using this technique to classify spam (Paul Graham) has a better explanation of Naive Bayes on this web page.

I've written my own naive Bayes classifier to identify spam, with less positive results than he reports. However, naive Bayes can be a very effective technique, and I can believe his results.

The two things you have to beware of when using it are "smoothing" probabilities of words you've never seen (you don't want them to always be zero, as straight naive Bayes will give you), and you need LOTS of training data for naive Bayes to work well. That means that you need to already have a fair amount of spam to identify spam well.

You can see a paper I wrote on using naive Bayes to classify hard drive failures here, or look for more stuff on naive Bayes on Google. Also, don't reinvent the wheel: Andrew McCallum has written a very good toolkit for doing these sorts of things in Bow.

But what's even worse, imho, is the horribly bad implementation. They encrypted only the file contents; file name, size and (what were they thinking?) the CRC were all in the clear. If you were using encryption to hide the fact that you possess a file you're not meant to, Pkzip will do you in real nice.

All in all an excellent example of how crypto works not.

Alex

• http://www.comp.glam.ac.uk/~NRHM/volume4/a-1998-9. htm
• http://citeseer.nj.nec.com/460740.html
• http://pespmc1.vub.ac.be/LEARNWEB.html
• http://www.dlib.org/dlib/june02/bollen/06bollen.ht ml

What about network shares in Windows or remote mounts in *nix? These are filesystem-independent, so unique identifiers like inodes aren't available.