Slashdot Mirror

They are talking about passwords entered after you log in. To things like sudo, or SQL front ends, or other crap like that.

I do not believe you are correct, from their document it appears they are focussing on the initial login. It is much easier as they know where in the stream to look. Trying to capture a sudo, etc. would be very difficult as you would need to analyze the entire session stream (that is almost certainly intermixed with writing emails, scripting, coding, etc.) and look for repeating patterns (and that assumes you are entering the password alot for various things).

Or loging into another machine.

Yes, they do address that. But now that involves them having a very, very (almost too) convenient location from which to sniff, twice (or they are sniffing from your machine itself). Using switched networks (super common) and decent physical security will make this really hard. Also, if you use RSA style auth on this operation too, you have circumvented it again.

I think the original posters statement that RSA authentication is the way to short-circuit this potential vulnerability is correct. Perhaps the use of the word "nullify" is a bit strong, but it sure makes an extremely difficult task even more difficult (perhaps approaching impossible). I have to be honest, I have always found RSA auth annoying, but this vulnerability has perhaps convinced me that it is the right way to go.

You might know Clifford Stoll, he is famous from the early days of the Internet when he tracked down a computer spy. And he's programming computers since the mid 60s.

But his reasons for keeping computers out of the shools are compelling:

• computers and educational software are expensive. Much more expensive than say good school books. And computers will break faster than books.
• computers are also expensive because they bind ressources which could be used better. E.g. teachers have to do system administration instead of teaching. And computers and educational software get old soon.
• educational software e.g. in physics only simulates reality. To understand magnetism you have to hold the real thing in your hand.
• and most of all he thinks that computers are easy to use and you don't need computers in school to become computer literate. You can learn how to use a computer (which is what people need) in only a couple of weeks.

After I read the book, I was convinced, that we should have computers in school, but only where they are really useful. Giving a laptop to every pupil seems much too expensive and the money should better be spent for conventional education (more teachers, better books, better libraries).

Clifford Stoll sometimes sounds too extreme or even fanatic to me, but then he has a lot of facts which prove his ideas. So you might want to read his book or give it to your school :-).

Linux World Expo is next week at the Moscone Center in San Francisco!

It's pretty cool to see how far Linux has actually come since the 'ol days. Maybe one day we'll finally settle this whole Gnome vs KDE stuff once and for all.... ;)

Speaking of parties..don't forget this party to raise funds for Dimitry's legal defense.

The plates in the fault system(s) around California are pretty much all horizontal and not vertical. See here. So it won't fall into the ocean, but it might relocate parts of itself. I researched this a little, as earthquakes are one of the reasons I don't want to move to California (Aside from crazy people, mudslides, fires, high taxes, and now the market conditions.)

I wondered if it was as bad as I thought. And in this case at least, the rumors aren't true.

BITTER
* 2001-08-22 16:19:50 ssh vulnerable to keystroke latency attack? (articles,encryption) (rejected)
/BITTER

BITTER
* 2001-08-22 16:19:50 ssh vulnerable to keystroke latency attack? (articles,encryption) (rejected)
/BITTER

Some folks in this thread have already mentioned SETI@Home, but I thought I'd point out that the collective (albeit highly specialized) processing power of the current SETI@Home array is twice that of ASCI White.

Researchers at The University of California - Berkeley have also done some research in this area. The article is here.

On many computers, greater precision operations do not take any longer than lesser precision operations, so it makes numerical sense to use the greatest precision available for internal temporaries. The philosophy is not to dumb down the language to the lowest common hardware denominator, but to enable the exploitation of the best capabilities of target hardware.

For floating point operations and expression intermediate values, a greater precision can be used than the type of the expression. Only the minimum precision is set by the types of the operands, not the maximum.

If you get different answers on different computers due to different roundoff errors, your software becomes unreliable. It's true!

People get confused by Intel's 80-bit FP arithmetic. Yes, the FPU expends some effort in rounding the 80-bit result back to 64 bits, but the result is not more accurate than a 64-bit FPU. In fact the answers will be exactly the same--this is mandated by the standard.

Anyone using floating-point arithmetic for anything serious needs to know exactly what the arithmetic model is. If Walter pursues this philosophy with his new language, he will make it unusable for numerical applications.

Walter needs to read:

David Goldberg, "What Every Computer Scientist Needs To Know About Floating Point Arithmetic," ACM Computing Surveys, vol. 23, pp. 5-48, 1991.

I could not find a copy online, but here is an interview with William Kahan, the Turing award winner who co-developed the IEEE 754 floating-point standard. Language designers should notice that Kahan implicates of Java and Fortran at the end of the article.

Yeah that's good. But what if the worm spreads fast, not like the lumbering Code Red random pot shot approach. This guy's theoritical Warhol worm thinks it could do it faster. Now that's scary. Let's just hope this one doesn't end up like the Morris worm. I really hope these guys who write these worms ,for just pure research, take the same care that biohazard researches use. Because these contaminants would spread at light speed.

I know what would get worms back into the media for a long time - a Warhol Worm. You want to read something scary about worms, go read that. Be sure to read the section "A Worst Case Warhol Worm". It gives me the shivers to think about it.

From the article: "A worst case Warhol Worm is truly frightening, capable of doing many billions of dollars in real damage and disruption. Since it can achieve complete spread in well under an hour, and could begin doing damage immediately on infecting a machine, human mediated responses offer almost no hope of stopping it. "

Complete spread in under an hour! Total destruction of infected servers!

Whee!

Watch for one of these coming out with the next major IIS exploit.

...orbit at an altitude of 35785 km ...

Using your figures of a 23 hour, 56 minute day and 35785 km the speed (note; not the same as velocity) of the satellite =
2 x pi x (35785km + radius of Earth) / 23.933 h

The radius of the Earth is about 6370 km (see here for info), so the speed of the satellite is:
2 x pi x 42115 km / 23.933 h = 11067 km/h

Now, suppose we put a satellite at, let's say, 43630 km from the surface. Gravity still works at this height. So you can still put an object in orbit. And suppose we set it moving in an orbit parallel to the lower satellite at a speed of, ooooh, about, 13127 km/h. What then?

Geosynchronous orbit is simply keeping the satellite over a particular point on the ground (as I'm sure you're aware). Or in this case we simply need to see if we can keep pace with the lower satellite (using your figures).

Let's think of an analogy - two runners going round a circular circuit. One runner gets an inside lane, and the other gets the outside lane. But the outside lane is longer (the radius is bigger) than the inside lane, so if the outside runner wants to keep pace with the inside runner he has to run faster. The speed he has to run depends on the circumference of his track - the ratio of his speed to the inner runner must be the same as the ratio of his track length to the inner track length. And because it's a circle, we can simplify that ratio to Radius(outer):Radius(inner).

Now, let's return to our satellites - suppose yours and mine start above the same point on the Earth, and we know that yours is geostationary. If mine manages to travel faster than yours around it's larger orbit then it will keep pace with yours and stay geostationary.

The ratio of distances from the centre of mass of the Earth is (43630km + radiusEarth):(35785km + radiusEarth) = 50000:42115 = 1.187:1

So my satellite needs to be moving 1.187 times faster than yours - 11067km/h x 1.187 = 13136km/h, which is pretty damn close to the speed I specified above (accounting for rounding errors). Why, it's almost like I cheated.

Essentially, a geosynchronous orbit can be achieved at any altitude, but it must be over the Equator (as I already explained to someone else) and the speed that the satellite must move is directly proportional to the altitude.

Please don't believe everything you find on Google - try doing the maths yourself first.

• 2001-08-11 13:18:46 Warhol Worm proposed: 15 minutes to total infection! (articles,bug) (rejected)

Here's the scoop (more meat at K5):

According to an article in the latest issue of the RISKS digest, Nicholas Weaver of UC Berkeley has written a description of a new type of worm, the Warhol Worm. He believes that using a divide-and-conquer method, all vulnerable machines over the entire IPv4 addressspace could be compromised in only 15 minutes!

`In the future, everybody will have 15 minutes of fame' -Andy Warhol

Too bad this is old news fellas. A group from UC-Berkeley has done an even more in-depth research project about the (in)security of wep, and can be viewed here:

Wep (in)Security

One of the important things to point out is that in the paper done by this group of people is that the also included active attacks, which is a pretty neat tool. I won't elaborate too much on this, but it is possible for a hacker (bad context) to act like a man-in-the-middle attack, sniffing your packets off the air, then doing whatever to them, then sending them to you (as if nothing every happened).

The sad thing is that most people don't even know that encryption is available on some of these models.

One other important thing to point out with wireless LANs is the new thing with war driving (similar to war dialing). What this consists mainly of is someone sitting outside in your parking lot and just surfing the net for free. There are also more complex stuff that is done out there, specifically in San Franscisco where the whole city was marked out by the http://www.dis.org guys, containing all the wireless LANs available as well as their SSID's (think of identification).

Here are some links on wardriving:
Mobile Wardriving
San Fran War Driving
General War Driving Info

One last thing to point out is that new technology that is coming out allows you to make a mobile sniffer device just using a Compaq iPaq, a Lucent wireless LAN PC Card, and a few other items (depending how sophisticated you want to get), and all of this can be done for under 1000 US dollars.

God bless Al Gore for creating the Internet.

And here is a link to their paper and additional information ... it would sure be fun to compare those for "similarities" ... ;)

I don't know. I think the whole issue here (and in other cases that have been reported by slashdot and a bunch of other places) shouldn't really be this law or that law -- it should be the entire system of law.

I believe the US as a nation has 'lost the game', as far as its legal system is concerned. If one merely takes a step back and looks at the situation, it is evident for what it is: a war of resources. It's like a real battle. Only, instead of planes and missiles, the two sides are launching lawyers at each other.

Now, I will not say 'I am not a lawyer', because half the time (former) lawyers are the ones that pass those abominable laws; the other half they are the same people that can't (or won't) interpret them properly. But what do I see? I see a country that has lost its sense of justice only to replace it with a 'just facade'.

In ancient times, lawgivers were respected people whose reputations preceded them. Entire cultures and city-states appealed to them and practically begged them to order their society. The ones whose names echo throughout history were wise men, that cared nothing for power or rule. And whose sense of justice was unmatched. Ask yourself: Who makes the laws of America today? And who finances them? Easy answer, isn't it?

Another point is that politicians are often the unwitting tools of other people's agendas. Not that politicians themselves aren't good at manipulating people, of course. But their arena is different. They are often clueless where subjects like - say - technology and IT are concerned. In their arrogance, they often don't even bother to research something before publically condeming it and pushing a bill that will further violate people's rights. All for the sake of sensationalism. And all the while, some shadowy person/corporation cackles in the background.

The Law should serve the citizen. Not the other way around. Maybe it's time for the few remaining wise men involved in law-making to look back into antiquity. At Lycurgus of Sparta. At Solon of Athens. Aristotle once said, on Solon's reforms: "Once master of the vote, the people became master of the constitution."

Get out of that maze. Or this will only be the beginning.

Pathway

http://www.law.berkeley.edu/journals/btlj/articles /08_1/McManis/html/text.html is a good read on Intellectual Property and Reverse Engineering law.

That is a good list of apps to use, but I would say that besides photoshop, only illustrator would be widely used or recognized by a majority of users. I cannot recall seeing any illustrator benchmarks, but I would imagine that the G4 does well enough for itself when manipulating the vectors. And powerpc chips have always had a good showing in the Seti@home arena as well. Check out the average CPU time per work unit.

As far as the 6 filter pshop bake-off went, it was done by TechTV, and the Pentium 4 and G4 each "won" an equal number of filter tests.

Those little Internet Applicances can surf the web pretty quickly, and they don't have even 500 MHz to play with. I think operating system overhead is a far bigger problem (one that plagues both Windows and Mac OS users)

I am fully aware that the Athlon can hold its own against the Pentium. My Athlon/Pentium comment was merely meant to show that there is some level of hypocrisy in the CPU wars...

-Renard

You'll notice that the chips do fall off immediately, but in the last few seconds (and the first few) of the clip you will see something in the top right of the view come in the shot very quickly then leave. This leads me to think that the shots were sped up intentionally to not bore the /.'ers, and keep your attenetion.

A big use for this kind of thing is for part feeding, that is a method of orienting parts on an assembly line. There is mention of this on Dr. Resnik's web page. Basically, you have the problem of presicely orienting a whole whack of (possibly complicated) parts as fast as you can to present them to the next stage in an assembly process.

Yes, a robot can accomplish this, but because of the motion that is involved they are slow, and because of the optical recognition involved they can bit, uh ... touchy :)

One method of dealing with part orientation is by things called bowl feeders. They are a bit hard to describe in type, but imagine that you have a big steel drum with a spiralling track up the inside of the drum. By vibrating the drum you can make the parts you are trying to orient climb up the track (beleive me, this sounds wierd, but it actually works). By changing the shape of the track you can force incorrectly oriented parts to fall back into the bowl, thus filtering out parts that are correctly oriented. So, only parts that are correctly oriented arrive at the top of the drum.

It's not quite as simple as that, but that's the general idea. Now, as well as this works (when it works), the problem is that whenever you change the shape of the part, you need to build a new bowl feeder! And building these things is not simple (or cheap).

I beleive what Dr. Reznik is trying to do with this experiment is not to prove that you can move poker chips around, but that you can build a programable solution to this problem; you can build one machine that will sort anything, given the correct programming of the controllers, thus alleviating the cost of prototyping things like bowl feeders.

So, to my mind, it's actually pretty spiffy, despite what others are saying here.

I know I've left out some details, and I certainly haven't researched Dr. Reznik's work, but hopefully I've given you (and some of the /. detractors) some idea where this work is probably headed.

The only thing I find a bit perplexing is, if he is proceeding towards the end that I've described, is how he is going to do this in three dimensions. Sure, he can rotate things in two dimensions, but what about more complicated parts?

I should also add that watching one of those bowl feeders in action is actually pretty creepy at first. Parts move up the track, but nothing else seems to be moving (the vibration rate is quite high, so you don't really see it). Mind you, they are freakin loud :)

Recognize what they're doing. They're vibrating a rigid plate in such a way that one object, out of many, moves. The system that drives the plate can produce both small rotations and translations. The vibrations have arbitrary waveform, and are generally asymmetric. It's the asymmetry that produces motion. That's all.

I could see this working for two objects, because you could vibrate the plate such that the center of rotation was under the one you didn't want to move, so it didn't go anywhere. But I had no idea how they make this work for N objects.

The novel result in the thesis is section 6.1. Figure out how a "jet" vibration works, and you'll understand the whole thing. The basic idea is that a rotational vibration centered on the point at which motion is desired is superimposed on a translational vibration in the desired direction of travel. When both vibration functions are suitably chosen, there's a very unexpected property: the feeding velocity is small everywhere except near the center of rotation.

This is counterintutive. You'd expect the rotational effects to be biggest far from the center of rotation, and zero at the center of rotation. Apparently the idea is that the forces induced by rotation interfere with the translational vibration that makes objects move. What puzzles me is that they're able to achieve zero feeding motion over most of the entire plane. But look at figure 6.2 in the thesis, showing the jet field.

That's really neat. But I don't get it intutively yet. Can anybody else explain it more clearly?

Artificial Intelligence - A Modern Approach by Stuart Russell and Peter Norvig is a great book that explains many concepts in AI. It is the book the most used in Universities around the world to teach this subject. It is not language specific, covers most aspects of AI, is okay for beginners, and goes in the details...

I really enjoyed this book and think it is a great buy.

There are more active things you can do. Follow the precedent of Henry David Thoreau and refuse to pay your taxes. It's called Civil Disobedience. Maybe you'll make it onto Letterman and have the chance to make a difference. :P

Seriously though, get out there and grow up. The U.S. has potential. "America -- Love it or Leave it" is supposed to be the catch phrase of the conservative assholes, not the intelligent next generation who could actually make a positive impact.

--netmouse

As a researcher, you do your research (your money and time)

Sorry, no. The money comes from grants or out of your institutions budget... That is *someone else's money*.

write it up in a suitable format for the journal you consider submitting it to (your time)

Time bought *and paid for* by somebody else.

[Snippage here, the pattern continues however.]

And your work is published in the prestigious journal, of which you will need a subscription (quite expensive) to view the results.

Our course it costs money to publish the journal, *somebody else's money* for which they rightly expect recompense. This cost analysis of two mathematics journals does not show all that great a return on investment.

Summing it up, the researchers spend a lot of time, money and good-will on the publications, whereas the involvement of the journal publisher is not that great after all.

Summing it up, someone else paid you to spend that time to produce the research and to publish the same. None of your money is at risk. Publishing a specialized journal for a small audience costs real money, and takes real work.

I work in the molecular biology field myself (which the article is relating to) and we have often jokingly considered opening up a journal, since this is a way to make money without much effort

You just might give it a try. It's a *lot* of work to publish a serious magazine that will be taken seriously. Until you are established, you'll have to work awful hard to get articles.Even after you are established editing takes a lot of work and time. Working with the printers takes time. Arranging for mailing takes time. Handling the (financial) books takes time. Dealing with customers and vendors and writers takes time. And those who expend the time expect to be reasonably recompensed for it.(Funny how you expect to make money at it, yet don't seem to think it fair for the already existing journal publishers to make money.)

... everything is done and paid for by others. While I am sure journals aren't exactly pots of gold, the distribution of who does what and who pays for what is a little odd.

Hmm.. You do the work of the research, and get paid for it, the publisher does the work of publishing and gets paid for it. I see no oddness there.

Rob Kirby has a list comparing mathematics journal prices. (Although some of the data is somewhat ('97) old now.) In particular, there is an excellent comparison of the (estimated) costs involved of producing the Pacific Journal of Mathematics (published by UCI) and Inventiones Mathematicae (published by Springer-Verlag).

I think most people in the mathematics community are aware of the issues here. However, certainly in the UK, there are far more pressing factors that determine which journals one submits to. Firstly, there's the obvious implications for one's career and promotion prospects: a paper in Annals of Mathematics or Inventiones looks better on a C.V. than a paper in some random journal no-one's heard of. Secondly, in the UK each member of staff in every university department is assessed for the quality of its research output (this is the dreaded Research Assessment Exercise). The ranking of each department (on the slightly bizarre scale of 1,2,3a,3b,4,5,5*) determines the level of research income that the government will provide. The quality of research is primarily assessed by each member of staff submitting their 4 best published papers over the last 5 years to be peer-reviewed by the RAE panel. They obviously can't read every paper, and so it's important for each department to submit as many papers in as many prestigous journals as possible: too few and your department could be closed down! Each submitted paper has to be put into the correct category: peer-reviewed journal, refereed conference proceedings, non-peer-refereed proceedings, etc. What is interesting is that refereed electronic journals are counted separately to traditional refereed print journals. The RAE panel claims that the two will be treated equally, but given that the panels usually comprise of the more senior figures in the discipline (hence, older and more adverse to innovations like electronic journals), and given the huge sums of money involved, it's regarded as foolish to submit papers to purely electronic journals in case they are perceived as being of lower quality.

I - and many others - would much prefer to publish in electronic journals, or to journals which are published by the academic organisations, but the potential career and financial risks are just too great.

200MHz is a lot. 1GHz is utterly ridiculous. The biggest concern with having faster chips is reducing battery life. Most, if not all, WinCE devices have a max battery life in the hours. Granted, most such devices have colour screens, but it would be foolish to say the faster chips doesn't play a part in it.

The Motorola MC68328 DragonBall microprocessor uses about 55mW. The power consumption for a modern colour LCD screen is about 25mW. So CPU power usage is an important factor in battery life. There is a paper on power consumption which is worth a read which shows that power consumption for the Palm Pro 16MHz uses between 26mW (sleeping) and 160mW for intensive tasks.

Now the ARM family has been designed for high MIPS/Watt ratios. Looking at the ARM10 tech specs the ARM chip uses 275mW when running, and can drop down to less than 5mW when idling, and in sleep mode uses 0mW with the option for fast wake up.

So what does this mean for the average Palm user? If the CPU goes to sleep in between uses, and idle mode is invoked during any period of inactivity, the new Palms should maintain good battery life while being a lot snappier to use. To me, it looks like you should get the same sort of battery life if you use it for the same sort of tasks. If you want to play MP3's all day long, then your Palm will be chewing around 350mW out of the batteries. Given that a single AAA battery gives roughly 1000mAH (milliAmp Hours) so that would imply a pair of AAAs would give you about 8 hours of continuous 100% CPU usage.

Cheers,

Toby Haynes

Also, a search engine (Cheshire2) running over the XML with a Very simple interface/display is available at:
http://www.o-r-g.org/~cheshire/osd/

Enjoy =)

-- Azaroth

-B

If the best tech always won we would all have been using Macs and Amigas by 87. I disagree, I've used both Macs and Amigas and they certainly aren't the best tech. Neither is Java. Compare Java to Python. Both are interpreted and cross-platform but Python is 10 times faster. I'm not saying the x86 is the best architecture(no one will ever need more than 640k of memory), it certainly isn't nor am I saying the best architecture included that various Motorola chipsets that plagued the Macintosh over the years, they have problems of their own. In the open market, the best advertised architecture with lowest price tag usually wins if it gives them everything they think they need. Java is in the same boat. Java is well advertised but does not give everything people need(speed.) Their are literally millions of theories to compress large random files into much smaller pieces(less than 50% compress but greater then 20%.) The problem is these types of compression of random variable would take so long to calculate, it would be the equivalent of 20 or 30 Seti Units. Java is easy, I've did some basic programminng(not enough to call myself an expert but enough to get by) but it is not BASIC or atleast as easy as VB. Plus, the little guys can only so far with the big guys on stepping on their tails all the time. This is why Java will never become BASIC.
----

As for my contention about the (non-)feasibility of AI with current technology, it's impossible to prove a negative. The burden is on the other side to prove that it is possible. I have yet to see examples or evidence based on current technology of true intelligence of the sort that Katz says we should be worried about.

None of your examples are evidence enough for me:

• chess--chess playing programs work in a single context (all they can do is play chess) and they work by heavy number-crunching and calculations. They don't understand the game, aspire to anything outside the game, or realize they are even playing a game.
• speech recognition, translation, summarizing business documents--these are basically computational/algorithmic parlor tricks. There is no understanding. MS Word doesn't understand the business document. It just calculates which words are the most frequent etc.
• conjecture/prove mathematical theorems--I'm not familiar enough with this field to comment on this one
• promising projects that might bear fruit--Things that "might" work are a far cry from things that "do work". At the moment, these projects, while very interesting, don't demonstrate what I would call intelligence, nor has any of them produced some leap that I think might eventually lead to intelligence.

You can certainly disagree about what constitutes "intelligence"; like I said, there is a great deal of healthy debate about this. But I have yet to see evidence of anything that looks like the kind of A.I. I would worry might take over the world as Katz describes.

It isn't that hard to figure the expected death rate among red bull drinkers (expected death rate w/o Red Bull times % of population drinking Red Bull), and ask yourself, is what we are seeing higher than we'd expect? I'll bet it's not.

It would be very odd if no one who drank Red Bull ever died. But for some reason our culture always treats death as an annomoly, which must therefore have a proximate cause.

In the longterm, the per capita death toll is exactly 1.

-- MarkusQ

What I would like to see happen with beowulf research is for everybody to connect their clusters via the internet (the proverbial Beowulf of Beowulfs) and go for the Holy Grail - the simulation of consciousness within the human brain. Folks, we are already using clusters of computers to search for consciousness...that's what SETI@Home is. But space is big, and very empty, and our odds of finding ET are small. Let's start a public project to search for HAL instead. Here is what we know about the brain. Here is the place to scratch the surface on how we think it generates consciousness. We are geeks, we are hackers, we joke endlessly anout Beowulfs of Beowulfs...what are we waiting for?

quote: The FDL does preserve the "viral" nature of the GPL in a couple of ways.

GPL

quote: Tom Christiansen has noted the GPL could be called "viral".

is

quote: The way it does this is by insisting that the code and anything
derived from it is also released with the GPL licence. In some senses
it is 'viral' in nature and it is this that is central to many
people's objections.

Also, it's worth noting that the word 'derived' is a little too vague.
Does a library linked to a GPL'd program need to be GPL'd also? Does
a program running on a free operating system need to be GPL'd?

There's no clear, obvious answer for either of these with the current
version of the GPL. The new version (3) is intended to fix some of
these shortcomings, but it's viral nature will remain.

a

quote: Despite the meaning "virus" normally connotes, the viral aspect of the GNU General Public License -- known formally as "copyleft" -- is a tremendous benefit to free software developers and the community they support.

viral

quote: And the licence is "viral", preventing the combination of copyleft and proprietary code.

license!!!!!

quote: This is the 'viral clause' of GPL -- it compels anyone releasing software that incorporates copylefted code to use the GPL in their new release. The Free Software Foundation says: "you must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program [any program covered by this license] or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this license."16

...but this guy sure ain't no Ripley.

Check out the three and four "legged" robot section:

Biomemetic Walking Machines

The three legged robot uses simple solenoids to achieve directional and rotational control (talk about a cheap actuator), while the four legged "bug" uses a simple mechanical system and open loop design (ie, you could build one of these devices from Lego with zero sensors, and it would work) - makes me want to break out the Mindstorms set...

Worldcom - Generation Duh!

The pair of legs powered by a chainsaw engine is more of a walking wheelbarrow than an exoskeleton. More pics here.

--

The pair of legs powered by a chainsaw engine is more of a walking wheelbarrow than an exoskeleton. More pics here.

--

• the results from the public's participation are made public, not kept for private gain as was reported with some drug and genome distributed research projects. If they want to make money from it, then they should pay for the CPU time, and state what they're doing up front
• computers aren't left on simply to run such programs

• the results from the public's participation are made public, not kept for private gain as was reported with some drug and genome distributed research projects. If they want to make money from it, then they should pay for the CPU time, and state what they're doing up front
• computers aren't left on simply to run such programs

(First, a preface. I'm a member of a computational astrophysics research group. We have ported our codes to the kinds of hybir d architectures of the machines discussed here, and have benchmarked their performances. Moreover, we have previously run on vector machines, so we have a fair idea of the pros and cons of the two approaches.)

While zavyman points out the basic problems inherent in parallelizing any discretized numerical model, the problem in obtaining good performance on hybrid architectures like the IBM SP-2s and SGI Origins which currently top out the top 500 list goes much deeper.

First, these machines are built around a hybrid architecture. Each node has a few processors (typically between 4 and 16, depending on the model), which utilize shared memory. These nodes connect to one another via an internode interconnect, with relatively modest bandwidth.

While this hybrid architecture allows supercomputer manufacturers like IBM and SGI to scale into the thousands of processors, it also introduces a substantial complexity into building of high-performance codes. Ideally, one would like to run threads-based parallelization on each node, and MPI between nodes, though the reality is that most codes in use use only MPI.

One can get decent scalability (into the hundreds of processors) when one runs physical models with limited communication -- ie, which simulate hyperbolic PDES like those of hydrodynamics (as zavyman describes above). However, things become more interesting when one considers more varied physics, such as that involved in solving elliptic PDEs (such as Poisson's equation for self-gravity or electrostatics). Because elliptic equations connect everything with everything else on the spatial domain, the communication costs ARE MUCH HIGHER. It is extremely challenging to build a multiphysics code with such varied parallelization demands. Indeed, it is a fair statement that no one has yet achieved excellent performance on anything close fo the thousands of processors available on these hybrid machines. For instance, another poster describes a climate model available from another research group. However, if you dig deeper, you find that they state,

"ForesightWX uses an IBM 12-node system with 52 processors working 24 hours a day. The cluster fits snuggly in a small room. A decade ago the same power would have filled the building."

52 processors is a far cry from the thousands of processors available to the users of these machines. Since each processor is slower than a vector processor like the Cray (by about a factor of 3 - 5), and assuming ideal speedup, such modest levels of parallelization lead to speedups of about 10-15 relative to a single Cray T90 processor. It is quite evident that there is little net gain over running the same simulation on 8-16 T90 nodes.

Moreover, due to the hardware constraints described above, IT MAY VERY WELL BE THE CASE WE NEVER SEE EXCELLENT MULTIPHYSICS PERFORMANCE ON THEM.

(One can get better parallel performance by increasing the problem size, but as the article points out, doubling the resolution of a simulation increases the cost by a factor of 16; hence, simply increasing the problem size may lead to unacceptably long computation times.)

I think massively parallel architectures will ultmately be the wave of the future, but there is little getting around the fact that the current generation of IBM-SP2s are dogs in the performance category.

Bob

As pointed out by a previous poster, this author is a beginner in most of these languages. I had an interesting experience in a graduate level CS class here in Berkeley on optimizing a matrix-matrix multiply routine. (Interestingly enough, the class was on parallel computing -- the point of the exercise was to learn just HOW INCREDIBLY important the serial part of one's algorithm is, even when one has oodles of processors available).

The results are interesting for a number of reasons.

(1) The "naive" algorithm, even with optimization, performed at about 50 MFlops (the Suns we used had a theoretical peak of 333 MFlops).

(2) With excellent optimization, pulling out all the stops, and using all the tricks available (unrolling loops, deallocating pointers to local variables, etc.), teams of just two students working for a week could get EXCELLENT performances (ie, within 10-20% of theoretical peak), approaching those of Sun's built-in library, and exceeding those of some existing libraries (like PHiPAC).

(3) Different groups with different approaches got very widely disparate results -- some barely exceeding those of the naive algorithm.

In sum, how ones goes about coding an algorithm can make ENORMOUS differences in the performance of a code. This is particularly true with numerical algorithms in C and other languages with pointers, where some compilers have great trouble optimizing routines using pointers, since the values are not known at compilation time. Taking this into account, I wouldn't give this guy's results much credence at all.

However, with the help of a lot of experts in the various languages, it will be possible to get a much better appraisal of the relative performances of different languages. A close analogy exists with these benchmarks and the SPEC open standards evaluations for CPUs. The only fair way they found, to compare across all CPUs and compilers, was to allow a very strict non-optimal compilation, and no-holds barred compilation. The same is true here -- we need to get teams to go no-holds barred in the creation of the best possible codes for each language.

Bob

As pointed out by a previous poster, this author is a beginner in most of these languages. I had an interesting experience in a graduate level CS class here in Berkeley on optimizing a matrix-matrix multiply routine. (Interestingly enough, the class was on parallel computing -- the point of the exercise was to learn just HOW INCREDIBLY important the serial part of one's algorithm is, even when one has oodles of processors available).

The results are interesting for a number of reasons.

(1) The "naive" algorithm, even with optimization, performed at about 50 MFlops (the Suns we used had a theoretical peak of 333 MFlops).

(2) With excellent optimization, pulling out all the stops, and using all the tricks available (unrolling loops, deallocating pointers to local variables, etc.), teams of just two students working for a week could get EXCELLENT performances (ie, within 10-20% of theoretical peak), approaching those of Sun's built-in library, and exceeding those of some existing libraries (like PHiPAC).

(3) Different groups with different approaches got very widely disparate results -- some barely exceeding those of the naive algorithm.

In sum, how ones goes about coding an algorithm can make ENORMOUS differences in the performance of a code. This is particularly true with numerical algorithms in C and other languages with pointers, where some compilers have great trouble optimizing routines using pointers, since the values are not known at compilation time. Taking this into account, I wouldn't give this guy's results much credence at all.

However, with the help of a lot of experts in the various languages, it will be possible to get a much better appraisal of the relative performances of different languages. A close analogy exists with these benchmarks and the SPEC open standards evaluations for CPUs. The only fair way they found, to compare across all CPUs and compilers, was to allow a very strict non-optimal compilation, and no-holds barred compilation. The same is true here -- we need to get teams to go no-holds barred in the creation of the best possible codes for each language.

Bob

But I'm not sure how (if...?) this argument could be applied to the MS license under consideration. On one hand, their primary objective is to ensure that none of the SDK code gets "trapped" in a GPL program which is obligated to release its (and therefore MS's) source. This represents a poor understanding of the GPL, but given this interpretation, this could be a reasonably reponse.

In a broader sense, though, as the yahoo article and all of us seem to be arguing, this is an attempt to squelch the use of open source software by outright banning its use. However, it seems about as silly as a license saying "MS Word may not be used to type an article which disparages Microsoft or its products," or "you may not use Sprint long distance minutes to advise others not to use Sprint long distance." I mean, those provisions wouldn't be legal, would they?

Maybe a better comparison is to the right to reverse engineer. Courts have ruled (tons of citations in this article) that it's legal to reverse engineer, for example, video game systems, even though the manufacturer obviously objects. Basically, it seems like IP owners, including microsoft, put in tons of garbage restraints in their licenses in an attempt to expand their IP rights by essentially seeing what gets through.

Also, as per that article, this could be seen as "a lockout device to restrain competition," which would make it illegal.

---

To astronomers, flat means that the usual rules of geometry are observed - light not being bent by gravity travels in straight lines, not curves.

Yes, the article was misleading on this point.

A corrected version of the BBC sentence would be:

"light not being bent more than an average amount by gravity travels in straight lines, not curves."

The experiment was about real light, not light in some hypothetical 'empty' space.

For astronomers, a flat universe also means that the negative curvature of geodesics being pulled apart by the expansion is exactly balanced by the positive curvature from gravitational attraction (and the expansion continues forever, but with no exponential runaway).

You might like to look at this press release from Berkeley:

• MAXIMA, a balloon-borne experiment directed by UC Berkeley, finds evidence for a flat universe, inflation and a cosmological constant.

Pooh. That's nothing. I can make my Palm mimic a Unix desktop, or even <shudder> a Windows desktop! Link: here

Wasn't the entire alpha design team head hunted over to AMD? I think that is why the athlon uses the EV6 bus. The original alpha chip, designed by the guys that are now at AMD, is still the fastest processor on the planet. Seen their Seti
results? 59 minutes to do what takes my 1.4Ghz athlon 1 hour 50 minutes to do (old client, ars's benchmark unit).

I agree that single sign on is something that we really need to work on. That's why I started the Webwide discussion group to try and hash out an open standard and some open source implementations. We haven't seen much action yet, but you're welcome to join.

As for storage, something like OceanStore would probably be best, but no full implementation exists yet, to my knowledge.

So let's get cracking!

About 18 percent of all the "dark matter" in the universe may now be made up of neutrinos. Anyone care to elucidate on this part?

There is a lot more mass in the universe than what we can see (what you think of as normal matter). The theory is that there is a bunch of dark matter that we can't "see". Since previously mass-less particles now have mass, this accounts for much of this mass we previously couldn't see and which we assumed was composed of dark matter.

On reading this I became curious what my own institution's policies were, so I looked them up. FWIW:

First of all, my school maintains separate policies for copyright and patents. The copyright policy states that copyright of student works resides with the student. The patent policy makes no specific references to students, but says that anybody who uses university facilities or receives money from the uni (including gifts and grants) has to surrender patent rights to the uni, though they get a chunk of any royalties.

Since different aspects of software can fall under copyright or patent law, sounds like this must be mighty confusing for the CS folks. My best guess about what this means is that if you invent something, the university owns the invention, but you own whatever you say about the invention (and in the case of software, you own the specific code you wrote to implement the invention).