Slashdot Mirror

You need to ask whoever it is that is administrating the web site at the FBI why there isn't source code available.

I got this link today out of my SANS newsletter. Dave, Marcus Ranum, and others developed their own scanners *and* provide C source code. Also, he has several reports on trin00, TFN, and stacheldrahtas well as pertinent links on the subject. http://www.staff.washington.edu/dittrich

I'd like to know more about the DoS that it looks for.

There was an extensive analysis of trinoo DoS networks on Bugtraq last month. You'll learn a lot more from Security Focus" that you will from the binary or its source.

Here are some and Trinoo links.

But, dosn't anyone realize that having the source makes it easier for the trinoo coders to see how they are being detected and then modify the clients?
Anomalous: inconsistent with or deviating from what is usual, normal, or expected

I'd like to know more about the DoS that it looks for.

There was an extensive analysis of trinoo DoS networks on Bugtraq last month. You'll learn a lot more from Security Focus" that you will from the binary or its source.

Here are some and Trinoo links.

But, dosn't anyone realize that having the source makes it easier for the trinoo coders to see how they are being detected and then modify the clients?
Anomalous: inconsistent with or deviating from what is usual, normal, or expected

This is not a Smurf attack, it is a distributed DoS attack and there are some large differences.

A Smurf attack is very simplistic, and easy to prevent your network from being used in a Smurf attack. The trinoo and TFN Distributed DoS attacks are not exploiting a misconfigured server, they are installing an Agent on a hacked machine (or hundreds of hacked machines) and controlling them via a handler tool.

The architecture looks like this:

&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp &nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp/---A gent
&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp---Handler---Ag ent
Client---Handler---Agent
Client---Handler---Agent
&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp---Handler---Ag ent
&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp &nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp\---A gent

So the Client controls the Handlers which in turn each have a myriad of hacked servers with Agents installed that it connects to. Depending on which version of the attack is being used (trinoo vs TFN or another of the 4 or 5 versions that have popped up over the past few months) different communication methods will be used.

The TFN uses some encyption (Blowfish) for it's communication between Client and Handler, which goes across 16660/TCP by default. It uses ICMP for it's communication unlike its predecessor trinoo which used UDP.

Communication from Handler to Agent is done across 65000/TCP, ICMP ECHO_REPLY. If your system is hacked and being used in a Distributed DoS attack then I recommend blocking ECHO and ECHO_REPLY.

From what I understand most systems being used in this attack are Solaris 2.x machines, with some Linux being scattered in there for fun. The Makefile has rules in it for Linux and Solaris, the default being Linux. http://staff.washington.edu/ dittrich/misc/tfn.analysis

forge

Went over to CERT
They claim they've been finding a client called Stacheldraht on compromised hosts, sometimes with up to 100 connections to other compromised hosts.
This is consistent with security claims at Dave Dittrich's site at U Wash
Basically, someone uses known remote root exploits (lpr, named, ssh, to name a few recent ones) and compromises hosts. Then he synchronizes them to DoS some target from someplace very safe. One person can thus appear to be a few hundred clients all attacking some target simultaneously. By making a trivial change he could move his target.
This is NOT a large synchronized group of people. It is one or at most a few good crackers just having a good time, hardly believing how much damage they are doing so easily.
The report names linux and Solaris as the machine types with makefile rules defined in the program, and the program has only been seen on Solaris 2.* in the wild.
German for "barbed wire".

IMHO, this is a wakeup call for network administrators; *every* network should be agressively scanned by its administrator daily to look for open ports, and all unneeded services should be shut off. Without fail.

there were mentions of trojan attacks against sites which in turn were to be used to create distributed DOS attacks mentioned in security forums recently.
Here is an analysis of the stacheldraht distributed DOS attack.
Cheers
Al

i'm not a solaris user or an admin, however, this article by David Dittrich analyzes some source code for stacheldracht, which was found by a security investigator in a stolen account. apparently there were makefiles for both linux and solaris. it doesn't run as well on linux,though, which is probably why it has been found only on solaris systems. the article also compares and contrasts the different DoS attack tools, i.e. TFN, TFN2K, trinoo, and stacheldracht.

1) stacheldraht"
2) trinoo
3) tfn tribe flood network
4) tfn2k
5) Cert's denial of service tools
Useful?

1) stacheldraht"
2) trinoo
3) tfn tribe flood network
4) tfn2k
5) Cert's denial of service tools
Useful?

While there's no denying that C++ is a steaming heap of offal, ST is not the last word in OO languages.

I'd rate Self, Dylan, Cecil, and Common Lisp as OO languages with better designs than Smalltalk.

This is not to say ST is bad; in fact it's wicked cool. It's just not the last word in OO.

The "code morphing" layer is what makes Crusoe stand apart from the rest. It optimizes on the fly the instuction set it's running on the fly. This means that your aps will run faster and faster as it runs. This layer is what gives the Crusoe it's speed.

The only way "code morphing" could run faster than native code is by exploiting runtime information to perform optimizations that are not possible at compile time. In other words, self-modifying code that runs faster than static code.

This is plausible, but that doesn't mean there would be no performance benefit in compiling native code. Research on self-modifying code is not unique to Crusoe---it's a very active area of research, and there are two major kinds: JIT and dynamic compilation. JIT, which you're probably all familiar with from Java, involves translationg code (typically from a foreign instruction set) and performing optimizations at runtime; dynamic compilation involves "staging" code at compile time to modify itself in a disciplined manner at runtime. JITs and dynamic compilation are very different in the nature of optimizations they perform; one of the major differences is that because dynamic compilation performs its analysis at compile-time, it can theoretically perform much deeper and more sophisticated optimizations.

Crusoe does no staging (it can't: it executes fully precompiled code), so its optimizations operate under severe time constraints. Therefore, Crusoe's code morphing is likely to produce code optimality akin to that emitted by a JIT compilation system: shallower analysis, shallower optimizations. Which almost certainly makes Crusoe's "code morphing" worse than native staged dynamic compilation would be.

In summary: my point is that self-modifying native code that improves its performance at runtime is entirely possible without "code morphing". On the other hand, binary x86 compatibility is arguably Crusoe's major selling point, so there's not much impetus for them to bother encouraging any kind of native code compilation. Anyway, I get the impression that Crusoe's entire architecture would have to be revamped if they wanted to run native code so it's a moot point.

If you're thoroughly confused by now, try visiting the dynamic compilation project at the University of Washington for more information on dynamic compilation.

~k.lee

(BTW: this does not mean that Crusoe does not embody any technical innovations. In particular, the hardware support the chip provides for its runtime code translation is very interesting.)

there was an article posted in our campus paper that is a little more in-depth than the scifi site.

Define sufficient; some parts of Linux are more threaded than others. Why rework something if it isn't a bottleneck?


The thing is it is. Almost everything is a bottleneck - especially on TCP/IP if it's not threaded - this is especially important too as linux becomes more GUIish.


For what? And if so, where are all the sites out there using it :)

Well, Microsoft.com and a few other sites out there are using IIS5 (microsoft.com is the world's largest commercial site).

And the speed record was set with the university of washington.

blah

• Warp drive runs counter to the "talk only" assumptions of SETI
  
• An image showing the curvature of space in the region of the travelling warp
  
• A more energy-cheap way to make the same warping effect
  

Why? Because netscape sucks. Using netscape is almost enough to make me stop using the Internet, or maybe enough to make me stop using Linux. I don't believe Mozilla will be any better. Garbage in, garbage out.

Check out the output of my script that reruns netscape whenever it crashes here.

The retina has some significant processing power of it's own...Both edge and movement detection are performed in the rear-most layer of the retina as a pre-processing step before the information is moved up the optic nerve.

This preprocessed information only goes directly to the old subcortical vision systems. The primary visual cortex (which is where your "conscious" vision is) gets its data directly from the retina as a raw monochrome image plus low-resolution colour data, which is mixed in at a higher level, about where groups of edges and junctions get merged into identifiable surfaces.

...and you left out everything below the cerebrum...The optic nerve feeds into the LGN (the expanded form of which is difficult to spell), which performs a great deal of processing in it's own right--it was, after all, the primary visual system for millions of years--including further (much more refined) edge and movement detection, some object recognition, attention, etc.

I believe you are referring to the Lateral Geniculate Nucleus (this is off the top of my head so I can't guarantee the spelling). This has taken centre stage in subcortical processing studies lately as the old notion of the so-called "limbic system" as a complete and discrete subsystem has lost favour.

The subcortical visual processing pathways which evolved in our premammalian ancestors are an entirely separate issue from the conscious vision I was talking about in my previous post. The subcortical pathways are almost completely hardwired according to instructions in our genes, and they don't carry an awful lot of information: just very basic notification of possible threats (eg there's something moving towards you very quickly: predator!) and possible rewards (something moving away from you very quickly: prey!).

The result of that is forwarded to the cortex for the final processing we all know and love, which extends the capabilities of the retina and LGN greatly. However, by that stage the input has virtually no resemblance to the output from the retina, having been thoroughly digested by the lower visual systems.

You imply that the visual cortex gets all or most of its input from the LGN. You have it completely wrong I'm afraid. There isn't any room for doubt about this; the primary visual cortex is probably the most intensively studied region of the brain and the description I gave of cortical vision in my previous post is at least broadly correct. the data definitely comes direct from the retina.

The old subcortical sensory pathways evolved without having any substantial cortex to talk to. Their primary output is to subcortical switching centres like the thalamus, and from there on to affective systems such as motor areas in the cerebellum, autonomic systems in the brain stem regulating breathing, heart rate etc, and the adrenal response. These are relatively inflexible hardwired reflex systems - instinctive responses - originally developed for our simpler ancestors.

The way that the cortex has been overlaid on top of that is to provide a parallel (and almost completely independent) set of alternative pathways with their own (cortical) information processing. And, as we all know, these cortical pathways are the ones that make us intelligent, perceptive - and, dare I say it - conscious.

Now, there are secondary pathways swapping data between the subcortical and cortical systems, but these are very high level and, as you very conveniently pointed out:

One further interesting tidbit: the descending pathways, from the cortex to the LGN, have 10x the capacity of the ascending pathways. Make of that what you will.

In other words, these bandwidth-limited ascending pathways represent a pretty trifling quantity of information passed on to the visual cortex from the LGN. It's a highly condensed "risk analysis" passed upwards for mixing in with the detailed sensory picture provided by the cortex to give it some emotional coloration. It adds an "Aaaaargh!!" factor to the beautiful scene of that huge muscular golden lioness bounding towards you.

The descending pathways are broader because they represent the cortex's afferent output, which is the whole point of having a cortex in the first place. These pathways feed into the old reflex pathways to modulate instinctive behaviour with intelligence and experience. Some of them are used when you "steel" yourself to do something which takes courage, or whenever you exert any kind of self-control against your base animal urges. The ones coming from deep inside the visual cortex carry learned information, (in this case, patterns representing recognised elements of a scene, such as a face you know) via the LGN to various midbrain structures where it acquires an emotional context (the face of a loved one or of a hated enemy). Whereupon appropriate signals are sent out to the hypothalamus, the brain stem etc. to prepare the body for the appropriate physical response.

(Incidentally, there are a number of us who are trying to combat cerebral chauvinism in neuroscience, and you aren't helping ;-)

Well, what can I say? Cortex roolz! It's what makes us human.

Seriously, it's obvious that you're fairly well read but you have to careful not to identify too closely with certain researchers' narrow preoccupations.

Subcortical neurology explains most of the behaviour of lower animals. It also helps us to understand a lot of human behaviour; there's no doubt that it affects us deeply, we are still basically animals precisely because of it. However it cannot explain those parts of our experience and behaviour which make us different form other animals. Our conscious lives - our language, our appreciation for music, art, good food; our thirst for knowledge, our complex sexuality - these all belong in the information-rich cortex.

Please have a look at William H Calvin's web site, he has (I think all of) his books online there. His background is in spinal and cerebellar neurons but he has a brilliant theory of cortical processing which goes a long way to explaining how discrimination of complex stimuli, conscious and subconscious thought, memory, learning and language all work at a neuronal and cell-assembly level. I particularly recommend How Brains Think to start with.

Consciousness is not what it thinks it is
Thought exists only as an abstraction

You misunderstand me. Of course grammar exists but it is not a complete, consistent logical system like mathematics (is meant to be), it is completely invented, mostly by accident.

The result is that there are some phrase structures which you want to add to in order to complete the sentence but you can't do it without breaking the rules or generating a sentence of incomprehensible drivel.

Most people prefer to break the rules than spout drivel, so for complex sentences in the real world, grammar often breaks down.

BTW, It's obvious that there is an innate potential for grammar in the human brain but I don't agree with Chomsky that we are all born with the same basic grammatical structures hardwired. If you wonder how it is that so many of us end up sharing a similar meta-grammar (to coin a phrase) then you ought to read William H Calvin's book The Cerebral Code (yes, the whole thing is online, thanks Prof!). He shows at the end precisely how neural structures to support basic grammar could form spontaneously to enable thoughts about who did what to whom, and with what. The same structures are probably used to generate the word order when the thought is spoken.

You may have noticed that the higher apes (principally chimps and gorillas) used in language experiments have demonstrated the ability to form simple grammatical structures too. There were also reputedly some experiments with an African Grey parrot which demonstrated similar ability (but I've not often heard the work cited and don't know how reliable it is).

PS. If you like Calvin's book, his latest one Lingua ex Machina is all about the evolutionary development of language. Like all of his books this one's online too.


Consciousness is not what it thinks it is
Thought exists only as an abstraction

You misunderstand me. Of course grammar exists but it is not a complete, consistent logical system like mathematics (is meant to be), it is completely invented, mostly by accident.

The result is that there are some phrase structures which you want to add to in order to complete the sentence but you can't do it without breaking the rules or generating a sentence of incomprehensible drivel.

Most people prefer to break the rules than spout drivel, so for complex sentences in the real world, grammar often breaks down.

BTW, It's obvious that there is an innate potential for grammar in the human brain but I don't agree with Chomsky that we are all born with the same basic grammatical structures hardwired. If you wonder how it is that so many of us end up sharing a similar meta-grammar (to coin a phrase) then you ought to read William H Calvin's book The Cerebral Code (yes, the whole thing is online, thanks Prof!). He shows at the end precisely how neural structures to support basic grammar could form spontaneously to enable thoughts about who did what to whom, and with what. The same structures are probably used to generate the word order when the thought is spoken.

You may have noticed that the higher apes (principally chimps and gorillas) used in language experiments have demonstrated the ability to form simple grammatical structures too. There were also reputedly some experiments with an African Grey parrot which demonstrated similar ability (but I've not often heard the work cited and don't know how reliable it is).

PS. If you like Calvin's book, his latest one Lingua ex Machina is all about the evolutionary development of language. Like all of his books this one's online too.


Consciousness is not what it thinks it is
Thought exists only as an abstraction

sorry, here's the link to the information.

Aircraft2. jpg
This is a cross ambiguity showing multiple aircraft as observed by the radar in 10 seconds of data. The processing was done for maximum doppler resolution and the aircraft is travelling at -21.5 +/- 1.5 m/s. The clutter at zero doppler is produced by signal propagating over the cascade mountains and by scatter from Mt. Rainier. We have verified the scatter from Mt. Rainier by varying the antenna pointing.

See this article on a device that can be used for walking simulation. It involves concepts some concepts that have been mentioned in a lot of the articles above. They have built a proof-of-concept device. (Scroll down the article for an image.)
By the way, I feel this discussion is an utterly interesting example of what brainstorming and "thinking outside the box" is about. Great!

Actually, sometimes "The Future" is a self-fulfilling prophecy. The architect (I forget his name) who popularized rounded and wedge shapes on buildings with assorted protuberances (I actually first became acquainted with the shapes as the background of The Flash's visits to the future) had a strong influence on today's architects. Materials, functional requirements, and expense affects actual designs but the influence is visible.

Flying cars were popular in the future of the 1950's. The founder of Moller has been striving toward it ever since, and test flight of the four-seater Skycar prototype is scheduled this year.

Transit methods such as the Disney monorail have also been around for a long time. Personal Rapid Transit devices have assorted designs, of which Taxi2000 is my favorite. 3-4 passenger vehicles which take you from your station directly to your destination station. No schedules, and a tiny overhead track. These PRT technologies have actually greatly gained capabilities with microcomputers, as now automated guidance and control is much simpler and cheaper.

And I recently read of a cellular wrist phone...

Pine is a text based email program that handles HTML mail pretty well.

-Mike
--

And yes, pine does read HTML email.

(Pine Is Not Elm!)

Even nominally having a browser isn't really enough if netscape really thinks that we aren't important. Here is a script that I wrote a few days ago.

#!/usr/bin/perl
#
# Netscape really does suck. This is a script
# that runs netscape and when it crashes it
# runs it again

while( 1 )
{
$rc = 0xffff & system( "netscape" );
if( $rc == 0 )
{
exit 0;
}
# It crashed.
unlink "~/.netscape/lock";
system( "date >>~/.netscape-sucks" );
}

See the results (so far).

If you are reading this and saying "What the hell is that?", allow me to introduce you to the fucking coolest genre of music ever created -
www.breakbeat.co.uk

Also, check out some work I'm doing in the jungle arena -
My software synthesizer page

trey

First off, the existing computer-to-video-tape is something different... what I believe he wants here is to use an IEEE-1394 (Firewire) equipped digital camcorder that uses MiniDV video tape as a data storage device. There are a good number of reasons one would want to do so: One, these cameras are selling very well and a lot of us already have one. Two, they are quite cheap (with low-end cameras less than $1,000 and tapes less than $10). Three, the tapes hold a *lot* of data: the DV-format video stream is 25Mbps (not including audio, subcode, and error correction: with these 36Mbps) and tapes generally hold 90 minutes giving a tape capacity of roughly 12.5GB (more if you include audio) so that we have a data storage cost of less than $1/GB with a read/write rate of 150MB/min (more if you use audio). While I have not investigated reliability issues from this angle, these figures compare very favorably with current traditional tape backup solutions. Plus, the average current SCSI tape drive takes very poor movies! :-)

Camera control is standardized so complete computer control over the tape's transport is trivial through the same (100-400Mb/s, gigabit coming soon) Firewire channel that the data flows over. Firewire and DV-over-Firewire is actually a peer-to-peer sophisticated networking architecture and stack of protocols and is very cool and much superior to USB (including USB 2.0). I'm very surprised it doesn't show up more on Slashdot and that more people aren't working on it.

One would not implement this by making a sequence of pictures (frames) and sending them to tape through a usual DV codec: DV-format video is based on a quantized DCT (like JPEG) as is thus not lossless. However, one can format an arbitrary data stream to fit the same format as what comes out of a DV codec and pass that to the camera. Played as video, it would look like whatever (noise) but the tape backup system would, as before, not use a DV codec, it would just lift the data out of the stream directly. In other words, the communications channel to the camera is not raw video-- it is lossless transmission of (usually, but not necessarily, DV-encoded) data.

Some people have brought up concerns of reliability. As mentioned above, the DV tape format includes error correction (which is *not* error concealment, which is what you do when error correction fails-- that is included too, but is device-dependent and another layer up whereas the error correction is an integral part of the standard). I do not recall offhand to what typical rate (i.e., 1 uncorrectable bit per x bits read) this is supposed to work, but one of the points of DV is its lack of generational loss. In other words, even video people care about not losing any data. In my experience, data loss in DV tape (even with tape that has undergone many passes) in video applications is pretty minimal to non-existant.

Of course, this is not much of a reassurance. But that is no problem really, because there is nothing stopping your backup software from pumping its output through the error-correcting code of your choice, say your favorite variant of cross-interleaved Reed-Solomon. Just about any storage system at all can be made as reliable as you want by layering error-correcting codes on top of it. There's no reason hard drives couldn't be made without any error correction at all (and they do plenty-- today's average HD is correcting lots of ``errors'' all the time and this is the way it's designed to operate) except for the fact it makes for a cleaner and more efficient implementation to have it done in the HD and not your CPU. It's no big deal.

Actually, DV is so cool a lot of us are wondering how it escaped product development before the marketers/biz people/bean counters dumbed it down...

If I ever get any free time (chewing through 4 3" binders of Firewire/DV docs. takes a while even when you have nothing else to do) this will be one of the things I'll implement, but it'll have to wait until I finish a subcode- preserving footage-mover first...

--Shawn, Cokus@math.washington.edu .

One would not implement this by making a sequence of pictures (frames) and sending them to tape through a usual DV codec: DV-format video is based on a quantized DCT (like JPEG) as is thus not lossless. However, one can format an arbitrary data stream to fit the same format as what comes out of a DV codec and pass that to the camera. Played as video, it would look like whatever (noise) but the tape backup system would, as before, not use a DV codec, it would just lift the data out of the stream directly. In other words, the communications channel to the camera is not raw video-- it is lossless transmission of (usually, but not necessarily, DV-encoded) data.

Some people have brought up concerns of reliability. As mentioned above, the DV tape format includes error correction (which is *not* error concealment, which is what you do when error correction fails-- that is included too, but is device-dependent and another layer up whereas the error correction is an integral part of the standard). I do not recall offhand to what typical rate (i.e., 1 uncorrectable bit per x bits read) this is supposed to work, but one of the points of DV is its lack of generational loss. In other words, even video people care about not losing any data. In my experience, data loss in DV tape (even with tape that has undergone many passes) in video applications is pretty minimal to non-existant.

Of course, this is not much of a reassurance. But that is no problem really, because there is nothing stopping your backup software from pumping its output through the error-correcting code of your choice, say your favorite variant of cross-interleaved Reed-Solomon. Just about any storage system at all can be made as reliable as you want by layering error-correcting codes on top of it. There's no reason hard drives couldn't be made without any error correction at all (and they do plenty-- today's average HD is correcting lots of "errors" all the time and this is the way it's designed to operate) except for the fact it makes for a cleaner and more efficient implementation to have it done in the HD and not your CPU. It's no big deal.

Actually, DV is so cool a lot of us are wondering how it escaped product development before the marketers/biz people/bean counters dumbed it down...

If I ever get any free time (chewing through 4 3" binders of Firewire/DV docs. takes a while even when you have nothing else to do) this will be one of the things I'll implement, but it'll have to wait until I finish a subcode-preserving footage-mover first...

--Shawn, Cokus@math.washington.edu .

Something vaguely like what you suggest is described by John Cramer's The Transactional Interpretation of Quantum Mechanics. The collapse of the waveform is still real, but I think you may find it more palatable.

See Larry Snyder's work on ZPL at the University of Washington. The idea of embedding parallelism into the language is nothing new.

For a different take on uncertainty and related matters, I recommend reading The Transactional Interpetation of Quantum Mechanics by John Cramer at the University of Washington.

One of the things that this paper made me wonder about is the implications of the exitence of the present moment. It seems to me that the simplest reason for there being a present moment is that the future is in some sense indeterminate. Whether this means that we have free will or not is an open question (could be only God does or everything could be random and we are rats in the maze) but worth considering. At any rate, the fact that we can percieve the present moment to me argues strongly that we are active in it. If this is really the way that reality is structured, I would be very leery of someone claiming that evolution would not find it and make use of it.

The situation may even be more subtle than my poor description above. Suppose that the future is inteterminate but constrained in some manner. The present moment might not be a point but have some sort of width described by a Poisson-type weighting and different possible futures get to vote, but the further out you go the less say a future has. One of the possibilities opened up by this model is some variation on Teilhard de Chardain's infamous "Omega Point" teleological God.

This model of cognition may also allow us to appreciate the concept of infinity by doing an infinite number of back and forth transactions in a finite amount of time. Sort of what Penrose was trying to say, but with a somewhat different mechanism.

Anyway. I am a mathematician, not a philosopher or a physicist, (although I play one on TV, much to the chagrin of my brother the experimental physicist!)

It would be extremely difficult to target such a thing since they are basically created by the equivalent of impulse or point sources and propagate in a circle around the point of origin. Those who think that shape charges could be used to target tsunamis should sit in a bathtub and attempt to target a surface gravity wave by poking their finger or dropping their rubber ducky in the water. You can push a whole lot of water in one direction, but that would be an advective process that would propagate not nearly as far as a wave.

The damage done by tsunamis is due to their kinetic energy being transformed into potential energy when they near land. After generation, they travel at tremendous speeds through the ocean. The speed of propagation is basically SQRT(G*H) where G is the gravitational constant and H the depth of the water, i.e. hundreds of miles an hour in the deep ocean. Their surface signature, on the other hand, is lost in the noise of the other waves on the ocean surface.

When they near land and the seafloor shallows, the hundreds of miles an hour worth of kinetic energy is transformed into a huge wall of water that, while moving much more slowly than before, still packs a hell of a wallop. Another key factor is the general variation of the seafloor topography as it shallows. Refraction and reflection processes can focus energy such that what would be a 5-10 foot wave on a flatly sloping beach could be several times higher. This is basically the same process that makes some beaches much better for surfing since the offshore topography focuses the wave energy.

You could probably use an explosion to effectively generate a tsunami if the situation were just right, i.e. you have a seafloor configuration near your target favorable for refraction processes significantly increasing the size of your generated wave. Otherwise it would take one hell of a huge explosion to impart the same sort of energy you get from a massive seafloor shift or a large bolide impact, and that could end up doing as much damage to you as to your target.

As to the suggestion that additional waves can be generated to cancel out a tsunami, ponder (for just a second, Pinky) how you can cancel out a tsunami that has reached the shore and become a huge wall of water. At that point it's no longer a wave since nonlinear processes have turned turned the tsunami into an advective process. And if you attempt to cancel it out while its still a wave in the deep sea the counter-wave and the original will certainly sum to zero at some point, and then continue on in opposite directions to wreak havoc.

I'd recommend checking out the University of Washington's Tsunami site for further theoretical and historical information. On a side note, many of the foundations of modern oceanography were laid by the research performed by Walter Munk and others during WWII. Most of the work involved wave forecasting such that the wave environment during certain invasions could be predicted sufficiently accurately to avoid overly large surf conditions.

Good information can be found at the University of Washington's Research TV site. http://www.washington.edu/researchtv/

And where are the breakthru languages?

You just aren't looking.

• Cecil

• Dylan

• Haskell

• Objective CAML

• Sather

• Self

intercepting the solar wind [4 Mb, MPEG format];

the M2P2 Current system [8 Mb, MPEG format].

intercepting the solar wind [4 Mb, MPEG format];

the M2P2 Current system [8 Mb, MPEG format].

intercepting the solar wind [4 Mb, MPEG format];

the M2P2 Current system [8 Mb, MPEG format].

While they are at it, they should stop teaching The Theories of Relativity, The Church-Turing Thesis, etc. And isn't Newtonian Physics all but an approximation, anyway? And according to Heisenburg, can we be certain of *anything*? I propose we abolish the Education system until they figure out for sure exactly what they are teaching. Me

We will be able to use lasers to read the data off of these structures after all:

Light Microscopy: Resolution beyond the lightwave barrier

Hopefully this will quell the reliability issues raised elsewhere regarding these devices.

It will be practical to use these devices in very small ruggedized devices by use of an array of solid state lasers. This will allow very fast, highly parallelized reading.

The peer review process is, I think, very effective at picking out the good from the bad. Even if one wants to put all ones journals on the internet, you would still want peer review. Actually, the main reason is more self serving, in that you get a lot more credit for publishing in peer reviewed journals, and you need to publish if you want promotion, tenure, or salary raises.

That being said, the main problem with paper journals is that subscription costs are very high. Actually, in many cases a lot higher than one would think the costs are. There has been a lot of noise about this in the math community.

So switching to internet journals seems like a good idea. In fact, a several established journals are also putting their stuff on the internet, for example, the American Math Society, and those journals published by Springer Verlag. However, you still have to pay the subscription costs. (Springer Verlag's are particularly high.)

Recently, there have been new, internet only, journals. Two that I know of are The Electronic Journal of Probability, and The Electronic Journal of Differential Equations. But I know that there are a lot more. These still employ peer review to look at papers, and have the same standards as regular journals.

If you just want to put your paper on the internet for anyone to look at, there is the Mathematics Archive, which anyone can contribute to. (Actually, it is part of a much larger archive where there are physics papers, etc.) Many many mathematicians, including many prominant ones, put there papers there, as well as sending their papers to journals. This is really a preprint archive.

And of course, you can always make your own web page and put your papers there, just as I have at http://math.missouri.edu/~stephen/prep rints/. Actually, many mathematicians do this.

In the end, I think that this open source idea of publishing is important to reduce costs of publishing. But I don't think that we should get rid of peer review.

One of my CSE professors at the University of Washington worked with Dave Cutler at DEC. He said Cutler was the "best programmer in the world". He said Cutler would always have two piles of printouts on his desk: a very short stack of his code and a tall stack of all his test code.

Unfortunately, Cutler is now in his own little world, racing cars and working on a Win64 project (codename "Sundown" harhar) that will probably never stabilize..

I use Emacs as an IDE from the Cygwin compiler/tools on NT 4.0. MUCH easier going back and forth from Linux to Windows if your key bindings are the same. Check it out at http://www.cs.washington.edu/ homes/voelker/ntemacs.html

Ted Kaehler has some interesting comments on the excelent book "The Cerebral Code" (which is available online). Ted calculated the capacity of the typical human cortex at 500 terabytes. While I don't agree with the way he did it or with his result, it is worth lookin into this.

While I enjoyed this post and felt it was spot on most of the time, I would like to point out that there are more objective interpretations of the QM formalism than the Copenhagen Interpretation. For an interesting discussion of this and a different interpretation, check out the transational interpretation.

An interesting variation on FTL information transfer (not to mention a serious slam of the Copenhagen Interpretation) can be found here.

Two years ago, when I was in the Windows world, I used UltraEdit, which everyone trapped inside Windows OUGHT to check out. These days, I love Emacs.

Eight words: GNU Emacs for Windows NT and Windows 95/98.

The ticket info is here, though notice they spelled his name wrong.

--

... that when our cars all run on liquid nitrogen, that it will be cheaper to put a fast computer in your car than in your house?