Slashdot Mirror

Why the hell should they have to create it physically if they can model it with something like ideas, solidworks or inventor?

How about because it's possible to model something in a CAD program that can't possibly be actually built in the real world?

See the Kansas City Hyatt Regency skywalk disaster (1981, killed 114 people, injured 180, called "the worst structural disaster in US history", at least until the collapse of the WTC, where there were aggravating circumstances) for an example of something that couldn't be manufactured as designed, with catastrophic consequences. (Henry Petroski's analysis in "To Engineer is Human" details the problems with trying to build the thing as designed, this report covers the actual changes made - Fig. 9 is most relevant.)

I was being distracted from my studies by the computer. My solution? I got e-texts. For example, it was hard to sit down and crack open Nandris' Handbook of Old Church Slavonic, but with the University of Texas' online course, I can position a chat window over a blank portion of the screen and study and talk over IM at the same time. Or, I can keep it in one tab and go back and forth between it and the BBC News website. In fact, I'm amazed at home much I'm getting done of studying, socializing, and keeping up with the news. Computer addiction is keeping me more productive, not less. Granted, I'm in academia, a profession based on soaking up as much knowledge as possible, but there are still millions of people who must be benefitting as much as I am.

It isn't that difficult to put some magnetic shielding around the processor. MuMetal, for example, could work pretty well.

"From 1986 to 2000 central Antartic valleys cooled .7 C per decade with serious ecosystem damage from cold"

'Antartic climate cooling and terrestrial ecosystem response' Nature 415: 517-20

----

"Both satellite data and ground stations show slight cooling over the last 20 years."

'Variability and trends in ANtartic surface temperates from in situ and satellite infared measurements' Journal of CLimate, 13: 1674-96

----

"Side-looking radar measurements show West Antartic ice is increasing at 26.8 gigatons/yr. Reversing the melting trend of the last 6000 years"

'Positive mass balance of the Ross Ice Streams, West Antarticia' Science 295: 476-80

----

"During the last four interglacials, going back 420,000 years, the Earth was warmer than it is today."

'CLimate and atmospheric history of hte past 420,000 years from the Vostok Ice Core, Antartica' Nature 399: 429-36

----

"Less Antartic ice has melted today than occured furing the last interglacial"

'Radiocarbon constrains on ice sheet advance and retreat in the Weddell Sea, Antartica' Geology 27: 179-82

----

The Sahara has shrunk since 1980

'Africans go back to the land as plants reclaim the desert' New Scientist 175, 21 September 2002.

----

On the other hand sea level *is* rising, as it has been for the last 6000 years since the satart of the Holocene, about 10-20 cm every 100 years.

http://www.csr.utexas.edu/gmsl/main.html

----

Hell I could throw in stats and references about the decreases in tropical storm activity, but I think I've made my point enough.

I agree that some level of moderation would help, but I don't think that that would really help the polarisation issue. Again, taking the Abortion article as a hypothetical example, once things have been running for a while I strongly suspect you find that you have 2 wildly innaccurate polar opposite articles with a lot of "agreement" from all the supporters of the relative camps, and a certain amount of "disagreement" from the opposing camp - but in practice that would be limited: why bother dealing with or even looking at the completely bullshit pro-life/pro-choice article when you have a perfectly good pro-choice/pro-life article to edit? In the meantime you'll have the relatively content free neutral article getting strong "disagreement" from both sides, and very little in the way of "agreement" from anyone.

There have been a number of studies and articles on how individuation, and letting groups separate, can lead the groups to polarise, often to extremes well beyond the initial views of any of the individuals in the groups. By forcing editors to butt heads over a single common article you tend to get less of the differentiation of separate groups required to drive such polarisation. Once you let groups separate off they will generally get driven to the extremes and little attention will be payed to the other point of view as it is simply incomprehensible to the group at the other extreme ("Why bother trying to go through and detail false claims in the pro-choice/pro-life page? The whole page is nothing but an endless stream of false claims and it would take forever to refute them all properly.").

Finally here's Wikipedia on the subject.

Jedidiah.

The scientific inquiry into 'magnetars' http://solomon.as.utexas.edu/~duncan/magnetar.html seems like a better approch to truth-finding. Discover magazine has an article about them, but this web site looks more informative. Crazy theories about time aside, real observations of the universe suggest interesting things are still out there to discover.

"If they are just trying these techniques now, then paleontologists need to start visiting with other departments. The sciences have evolved so much over they last 20 years that in order to do any real work you have to associate with people outside your discipline."

I could make a joke about paleontologists and geologists moving very slowly, but it wouldn't really apply here.

Actually, this is far from the first work using confocal microscopy on fossils. The article carefully points out that this is the first time it has been applied to fossils this old (i.e. this deep in the Precambrian), which is probably right, but paleontologists have been using confocal microscopy on younger microfossils for years now, though it isn't as commonly applied as it probably could be (e.g., see this article on fossil dinoflagellates and this catalogue of re-imaged type specimens). They've also used serial sectioning techniques for decades, and having access to something 3D is much more versitile than the old way (which was to make a scaled physical model from the information in the slices). Various types of CT scanning techniques are used if the object is big enough. The Raman spectroscopy application is fairly new, but usually ordinary optical mineralogy thin sections provide compositional information. It is only when you want to do a non-destructive technique that it would become important.

"Now, with fossils, I assume that you can only see the very top layer, because light wont penetrate rock that far."

Actually, most of the common minerals involved in fossilization are variably transparent at that scale (quartz, calcite, etc.), sometimes almost crystal clear (no pun intended). Geologists of all types (not only paleontologists) regularly make "thin sections" of rock that are a standard 30 microns thick or so. Also, many organic-walled microfossils are entirely extracted from the rock (like the dinoflagellates mentioned above, or fossil pollen and spores), so you can immerse them in whatever media you like, including standard biological mounting media. Being derived from biological cells, they fall in the same sort of size ranges and are tens to hundreds of microns in size anyway.

Bottom line, confocal microscopy works great for many of them, and as these imaging techniques have progressed, paleontologists have been applying them. I'm sure there are a great many discoveries left to be made as the use of such techniques slowly becomes more pervasive.

You may appreciate this mathematical formulation of Occam's Razor.

Impressive!

Student Quota: "Disk Quotas: 500 MB" http://www.engr.utexas.edu/itg/students/home_dir.c fm

The Student quota is 50 MB as my school (UT Dallas).

I would appreciate if you didn't use state owned, state maintained servers, bandwidth, and infrastructure for personal use.

I would appreciate if you didn't use state owned, state maintained servers, bandwidth, and infrastructure for personal use.

I'm guessing you have already considered the relevant University of Texas System polices, the Office of General Council Ethics Standards, and the ITS Policies. Sorry, I work for another Texas university. :-) Universities tend to be generous and tolerant of a personal computer on their network so long as it does not interfere with your work, does not violate any laws or policies, and does not interfere in any way with the network or other computing systems.

With that in mind, know that you and only you are responsible for the security of your computer and that you will be held responsible for any undesireable activity coming from your computer. If someone were to manage to compromise your computer and then attempt to compromise other university systems, you will at least be held responsible for not securing your own system, if not held responsible for anything coming from your computer -- or through it. If you are quite certain that you can keep your computer secure, then by all means run your own server and learn as much as possible. It's best not to experiment with production university systems. Besides, one could argue that using university-owned systems for your own purposes is a violation of the ethics policy. However, using your personal computer on the university network is no different than any student using a laptop.

That's silly. By the Bayesian Occam's Razor, theological hypotheses are disfavored, once you take into account the space of all possible predictions (not just the resurrection or whatnot), because they fail to be specific in the vast majority of cases, and you end up diluting the prior. (An example of this is the Ikeda/Jefferys fine tuning argument.)

That's silly. By the Bayesian Occam's Razor, theological hypotheses are disfavored, once you take into account the space of all possible predictions (not just the resurrection or whatnot), because they fail to be specific in the vast majority of cases, and you end up diluting the prior. (An example of this is the Ikeda/Jefferys fine tuning argument.)

String theory by construction is unfalsifiable.

Here and in many other places.

Here's my mockup of a size comparison (using their dimensions). I'm guessing it gets the biggest boost from being thinner. Unfortunately, they didn't release any side shots. http://www.cs.utexas.edu/users/kdresner/DSLiteSize Comparison.jpg

I've compiled many kernels, I even compiled my own kernel for a beowful cluster I built at IAT. But building custom kernels packagable by the distribution you're running seems to have become more complicated in the last 2 years. I've been quite happy with the default power management available in Ubuntu (suspend and hibernate work just dandy on my laptop!) so I haven't taken the time to figure out how the distro maintainers want you to do it these days.

Your comment about him not having "gathered the necessary voter signatures by now rather than waiting until the last sixty days before the deadline" betrays your ignorance of Texas' ridiculous election law.

Southeastern New Mexico?

Just by geography, it is clear that YOU are not representative of the average American.

You mean, applications of neuroevolution for games? Seems the hot topic du jour. Finally, bots with complex, non-deterministic, evolved behaviour...!

The University of Texas did this: http://nn.cs.utexas.edu/keyword?neuroevolution , game at http://nerogame.org/ . First you train your bots, then you test them in battle.

First of all, md5crypt() was invented by Poul Henning-Kemp from FreeBSD project and was released under "THE BEER-WARE LICENSE". Linux distributions adopted it later on.

Second, MD5 algorithm and md5crypt, while related, are 2 different things. While MD5 algorithm is broken (in academic world, not yet in real usage), I don't think md5crypt is broken. If that is not satisfactory, there is blowfish crypt from OpenBSD project.

Third, the first Java port of md5crypt I'm aware of was from Ganymede project (http://tools.arlut.utexas.edu/gash2/):

http://tools.arlut.utexas.edu/cvsweb/cvsweb.cgi/ga nymede/src/md5/MD5Crypt.java?rev=1.12&content-type =text/x-cvsweb-markup

I was searching for Java port of md5crypt because I was on a project that uses tomcat, which doesn't really provide a satisfactory password hashing algorithm (they only provide MD5 hash). Fortunately, I found Ganymede's port. (Anyone can find a java port of blowfish crypt?)

Last, I think md5crypt() can be used in any situation where password authentication is required, not only when authenticating against UNIX/Linux system password.

First of all, md5crypt() was invented by Poul Henning-Kemp from FreeBSD project and was released under "THE BEER-WARE LICENSE". Linux distributions adopted it later on.

Second, MD5 algorithm and md5crypt, while related, are 2 different things. While MD5 algorithm is broken (in academic world, not yet in real usage), I don't think md5crypt is broken. If that is not satisfactory, there is blowfish crypt from OpenBSD project.

Third, the first Java port of md5crypt I'm aware of was from Ganymede project (http://tools.arlut.utexas.edu/gash2/):

http://tools.arlut.utexas.edu/cvsweb/cvsweb.cgi/ga nymede/src/md5/MD5Crypt.java?rev=1.12&content-type =text/x-cvsweb-markup

I was searching for Java port of md5crypt because I was on a project that uses tomcat, which doesn't really provide a satisfactory password hashing algorithm (they only provide MD5 hash). Fortunately, I found Ganymede's port. (Anyone can find a java port of blowfish crypt?)

Last, I think md5crypt() can be used in any situation where password authentication is required, not only when authenticating against UNIX/Linux system password.

And then there is the paper from 2005 written by my former advisor:
ftp://ftp.cs.utexas.edu/pub/qsim/papers/Kuipers-aa ai-05-rev.pdf

When I got my degree in astronomy, the graduating class was six ... all guys.

When I got my degree in astronomy, the graduating class was six ... all guys.

When I got my degree in astronomy, the graduating class was six ... all guys.

You are talking about EE people, talk about sincere science people.
Here is the definitive link: http://oea.cs.utexas.edu/pdfs/us-jobs-3.pdf

Yeah this is a single University's external affairs office's report; but google "Demand for CS students exceeds supply". It really is true! Your point is valid considering other fields though, in engineering; for e.g. electrical/electronics/aeronautical.

Good detective work. This story dates back to the 1880's. Try using Google sometime.
http://www.qsl.net/w5www/marfa.html
http://www.tsha.utexas.edu/handbook/online/article s/MM/lxm1.html

They don't handle foreign affairs, but focus on at home threats (Which made me question the need for Homeland Security, but anyway). The point is, their entire job is to monitor the US and US Citizens.

I'm honestly not sure how you could be more wrong. Instead of just guessing at what you think they do based on your own bizarre interpretation of the name of the agency, you could try actually looking up real information on what the role of the NSA is. From their website we see that their stated mission is a dual one, involving "Information Assurance" to protect US information, and "Foreign Signals Intelligence" to collect and process foreign communications. Feel free to actually read the executive order that defines what the NSA does. Spying on US citizens is precisely what the NSA is prohibited from doing, and handling of foreign intelligence is part of their mission statement.

Jedidiah.

I firmly believe that the rise we are currently seeing is indeed the start of an accelerating trend. One that human activity has very little impact on, or if any, has been slowing the rise thus far by emitting clouds of smoke thus keeping temperatures artificially low. Now that we are cleaning up our soot emissions the CO2, and natural process that drive the climate, are regaining their direction. As you noted, we are woefully unprepared for such an event. I would say that we desperately need to start preparing, clearing the low lying areas, etc., but I'm realist enough to know that there is no feasable way to get the huge fraction of the human race that lives on coast lines to change their residence prior to a devastating catastrophy.

As to your comments about peak oil, read the literature more closely and in depth. Yes, oil execs and employees have started falling all over themselves to validate peak oil theories, but apparently, only because they suddenly realized that the peak oil nuts were their best friends. If the oil monopoly (removing Hussein took out the major producer who was not part of the Saudi, Kuwait, Exxon, Shell, Chevron cartel. Russia was broken up to let the cartel take over production there, next in line is Chavez in Venzuela) can convince the world that we have reached peak oil production then they have an indisputable excuse for maintaining higher prices and vastly higher profits. The peak oil advocates love to point out that oil production is down worldwide in recent years. but isn't it interesting that in fact it is down by almost the same percentage at every cartel field world wide. These fields are not connected and shouldn't be suddenly linked in their very close decreases in output capacity. Especially since peak oil states that we can not increase production, not that it will decrease and everyone, including the peak oil fanatics, is very quick to reassure their investors that they have plenty of in ground reserves for X number of years. I propose that this is evidence of global collaboration to set and maintain a specific level of output, regardless of the capacity that could be acheived if it were maximized. In fact, if you research the geophysical scientific papers on new sources and increasing recovery from old sources or souces that were previously inaccessible, you find that not only is there more oil available and accesible today than there ever has been before, but there is far more oil than than we have ever used combined. I would post links to these, but strangely all of the pages I had bookmarked over the last 10 years have disappeared or been replaced by peak oil pages. However, you can find some clue in the investment brochures and independent scientific reports. Now undoubtably these new sources and recovery techniques will cost more, so we have seen the last of cheap oil, monopoly or not.

I became interested in the science and economics of this because of my family's oil wells in west Oklahoma. In the late 60's and early 70's, during U.S. peak production, our 16 wells were producing approx. $200,000 in royalties per month. By the late 80's they were producing about $13,000 per quarter. Basically the oil company explained that, at that time, it cost $32 per barrel to produce from Oklahoma wells, and $6 per barrel to import from the middle east. So all of the wells were placed in maintenance mode, which is basically just enough pumped to lubricate the mechanicals, and at that rate the holding tanks were only emptied once per quarter, instead of weekly or twice weekly. When my mom sold th

The answer, as always, is "it depends". I'm firmly inside the "right tool for the job" camp.

Manual memory management is not free. In some circumstances, it can be quite expensive. There is a group of programmers who are best described as "rabidly anti-GC". These people are almost all completely unaware of the costs that manual memory management can impose on your code.

A multi-threaded program, for example, can allocate memory from any arena, but it MUST return a block to the arena from whence it came, which can cause all sorts of difficult lock contention problems, making free() much more expensive than malloc(). (Ask anyone who has written high-performance memory-intensive multi-threaded programs.)

In some languages, like C, the situation is even worse. In structure-hungry programs, you can end up structuring your code around data lifetimes, which precludes you from using the most natural, maintainable and efficient algorithms. Garbage collection frees you from this, as the GCC people have discovered.

I do recommend reading Paul Wilson's excellent survey paper on the topic. It answers a lot of your questions, though it's by no means the final word.

for more technical info, see his site at the Texas Advanced Computing Center. pretty pictures and software tool downloads even.

for more technical info, see his site at the Texas Advanced Computing Center. pretty pictures and software tool downloads even.

for more technical info, see his site at the Texas Advanced Computing Center. pretty pictures and software tool downloads even.

I don't think Kazushige did the original post for this thread "No you idiots - it's not about GOTO statements", he is too polite. But it is correct! Here is more information about his approach and a question.

From Kazushige Goto's webpage http://www.tacc.utexas.edu/~kgoto/#overview [utexas.edu]

"During the last decade, a number of projects have pursued the high-performance implementation of matrix multiplication. Typically, these projects organize the computation around an "inner-kernel", C = trans(A) B + C, that keeps one of the operands in the L1 cache, while streaming parts of the other operands through that cache. Variants include approached that extend this principle to multiple levels of cache or that apply the same principle to the L2 cache while essentially ignoring the L1 cache. The purpose of the game is to optimally amortize the cost of moving data between memory layers.

Our approach is fundamentally different. It starts by observing that for current generation architectures, much of the overhead comes from Translation Look-aside Buffer (TLB) table misses. While the importance of caches is also taken into consideration, it is the minimization of such TLB misses that drives the approach. The result is a novel approach that achieves highly competitive performance on broad spectrum of current high-performance architectures."

So why have chip designers, software developers and academics been focused on impoving cache preformance (e.g. L1, L2, L3, ...) and not improving TLB and virtual memory performance?

From Kazushige Goto's webpage http://www.tacc.utexas.edu/~kgoto/#overview

"During the last decade, a number of projects have pursued the high-performance implementation of matrix multiplication. Typically, these projects organize the computation around an "inner-kernel", C = trans(A) B + C, that keeps one of the operands in the L1 cache, while streaming parts of the other operands through that cache. Variants include approached that extend this principle to multiple levels of cache or that apply the same principle to the L2 cache while essentially ignoring the L1 cache. The purpose of the game is to optimally amortize the cost of moving data between memory layers.

Our approach is fundamentally different. It starts by observing that for current generation architectures, much of the overhead comes from Translation Look-aside Buffer (TLB) table misses. While the importance of caches is also taken into consideration, it is the minimization of such TLB misses that drives the approach. The result is a novel approach that achieves highly competitive performance on broad spectrum of current high-performance architectures."

So why have chip designers, software developers and academics been focused on impoving cache preformance (e.g. L1, L2, L3, ...) and not improving TLB and virtual memory performance?

See here for more details! It's unfortunate that the urban legend made it into one of my group theory textbooks.

They say the evidence is overwhelming yet you seldom see any of this overwhelming evidence.

People like you (and I'm not trying to be rude or point fingers) always say there is "overwhelming evidence" but never actually provide it.

No-one disputes that life forms today are more complex than they were a few billion years ago. The question is how that happened.

You say it is valid to "extrapolate" microevolution to macroevolution, yet you provide no evidence of macroevolution.

First, let me say this about "macroevolution":

There is no difference between micro- and macroevolution except that genes between species usually diverge, while genes within species usually combine. The same processes that cause within-species evolution can be responsible for above-species evolution, except that the processes that cause speciation include things that cannot happen to lesser groups, such as the evolution of different sexual apparatus (because, by definition, once organisms cannot interbreed, they are different species).

In other words, the only border for distinguishing macroevolution from microevolution - the species line - is an arbitrary one made for the sake of species study, not a border based on some difference of process. The same processes proven to work on one side of this line can operate on the other side, and acheive the necessary results.

Once two lineages are reproductively isolated from each other, they evolve more and more differences that they share but the other lineages don't. This phenomenon works whether the division is sexual, geographical, or merely physical due to some behavioral quirk (for example, parasitic worms that prefer a particular animal may never interact with similar worms that infest another species of animal, even though the two animals share territory.) There's no reason these processes can't apply to all lineages, back to the first eukaryotic (nuclear) cell. Even the changes in the Cambrian explosion (which I'm sure you're just dying to mention) are of this kind, although some scientists additionally theorize that the gene structures of these early animals were not as tightly regulated as modern animals, and therefore had more freedom to change.

The process of "macroevolution" can be, and has been, explained via the same principles that cause inter-species variation to occur. As of yet, there's no need to postulate some additional, perhaps mythical, force that caused, for example, dolphins to end up with finger-bones in their flippers. If you choose to believe instead that "miracles" caused it, that's between you and Occam's Razor. Just the same as we accept that the law of gravity works on Pluto the same way as it does here - even though we have yet to send astronauts there to attempt to play basketball and find out "for sure" - we accept that the biological processes of evolution we have thoroughly documented and seen the trappings of today were just as effective ten billion years ago.

Which brings me to my second point. You bandy around those tired lines "evolution has gaps, it has not been proven, it relies on faith", but that old dog don't hunt no more. The supporting evidence, and the direct observed evidence, of speciation - that is, what you would call "macroevolution" - really is plentiful. We have moved beyond evidence that it "happened". We have actually seen it happen in the lab(*). For a nice thick sample, do peruse the slide collection for the University of Texas' Biology 304 class.

(*) Futuyma, _Evolutionary Biology_, 2nd edition, 1986:

In one case, finally, a new biological species has arisen spontaneously in a laboratory. A strain of _Drosophila_paulistorum_ when first collected was interfertile with other strains but developed hybrid sterility after being isolated in a separate culture for just a f

Did you have to use the Swedish Chef translator? :)

Dude... the Yuan dynasty? The largest empire the world has ever seen? Stretched all the way to Poland? Made the USSR look small in comparison?

Not a world superpower... Yeah, right.

Here's a map of China from 1300-1405. Yeah.

Didn't Shakespeare already write all the blockbuster plots?

Whoever owns him will be bigger than Elvis.

Uh oh...

The US put ICESat (Ice, Cloud, and land Elevation Satellite) into orbit in January, 2003. It uses laser ranging to map topography. While it has had its own difficulties, it has been collecting data.

We were looking forward to having CryoSat data to compliment what we've collected, for data verification, additional accuracy, and for extended lifespan of observations. This is a true loss.

For more information on ICESat:

http://icesat.gsfc.nasa.gov/
http://www.csr.utexas.edu/glas/

I have checked this out and still prefer gmailrss - something I rolled up using perl scripts, that sits on top of gmail and used the gmail UI to make it behave like an rss reader read. check it out

Actually TFA was a bit light. I got my info from better links people posted. Here's a good source.

-

I thought the point of TRIPS was to make the chip do all the scheduling (ie the Data Flow architecture) rather than depend on the compiler generated sequence of instructions.

A superscalar chip has dynamic placement and dynamic issue -- the hardware is responsible for deciding which ALU each instruction goes to, and when they get issued. Doing this well requires a lot of transistors and costs power, and the hardware is rediscovering inter-instruction dependencies that the compiler could have told it.

A VLIW chip has static placement and static issue -- the compiler is responsible for deciding which ALU each instruction goes to, and exactly when it gets issued. This is really hard to do well due to variable latencies of instructions, eg. due to cache misses. It puts too much of a responsibility onto the compiler.

TRIPS is static placement, dynamic issue. So the compiler schedules the instructions onto the ALU grid. But then the hardware issues the instructions when their inputs are ready. It's intended to find a good middle ground between superscalar and VLIW in terms of what the hardware has to do, and what the compiler has to do.

Read the IEEE Computer overview paper http://www.cs.utexas.edu/users/cart/trips/publicat ions/computer04.pdf from 2004 for a much better idea of how it all works.

The TRIPS homepage has nine published papers on how this design will work and a schematic diagram of what they're expecting the design to end up looking like. They are also promising simulators and compilers later this year.

And they're explicitly working on making it work well with plain old C code. (I'm still wondering, though, if it might be advantageous to use languages with explicit dataflow support, like Oz.)

the homepage for the TRIPS project: http://www.cs.utexas.edu/users/cart/trips/ because the article doesn't do a good job at explaining the idea, which I think is very interesting. It's not mere branch prediction these people are talking about, and it's more than dumb parallel processing. They are basically fragmenting programs into small dataflow networks.

I recommend you read this paper. It gives a great overall picture of what TRIPS is all about and is actually really cool. (I read it about a year ago).

I am an ECE grad student at UT Austin so I know quite well of TRIPS. In fact I often speak with Doug Burger himself because he's the faculty advisor for the UT Marathon team, of which I am a member. (By the way, his name is "Burger" not "Berger"). I think TRIPS is an awesome concept and its exactly the kind of project that I wanted to be a part of when I became a grad student at UT. I also know Steve Keckler because I'm taking his advanced computer architecture course this semester, and we're actually spending a good chunk of time talking about TRIPS (course schedule).