Slashdot Mirror

Where to begin... data clustering is very much an AI technique. I did my PhD dissertation on it, in an AI laboratory at a major research institution. That some techniques may be simple (such as k-means) does not mean that they fall outside of AI.

Clustering is also known as unsupervised learning, and is an essential tool for many other higher-level AI technologies, including machine vision.

The field of AI is very broad, broader than you probably think. It encompasses clustering. Pick up any good AI book or look at top conferences in the field and find out for yourself. AI has always had the problem that as soon as someone understands part of it, that part is no longer "intelligent" and is now just a "technique". By that logic, AI would not exist.

ANSI C isn't portable; it's only standardized.

Take a look at a wonderful list of C weirdness which is mandated by the ANSI C89 standard, but which is supported spottily across different platforms.

ANSI C doesn't even specify how large a char should be, for crying out loud. C is many things, but portable ain't one of them. Try taking a codebase written in ANSI C for an embedded microprocessor and compile it for Big Iron. Dollars to donuts says that somewhere in your codebase there are going to be implicit assumptions (about the size of a word, about how memory is allocated, etc.) which are going to be wildly platform-dependent.

ANSI C is standardized, which for some reason people confuse with being portable.

Assuming the computations come to the same conclusion.

But e.g. 96-bit extended floating point calculations on Intel/AMD hardware will generally give different answers than 128-bit Sparc or Altivec calculations, and I'm sure you encounter similar problems in the high-end graphics market.

Compare: Matlab's Loss is Nobody's Gain or How JAVA's Floating-Point Hurts Everyone Everywhere . (Scroll down to the bottom of the page; they're PDF files, which is why I didn't link to them directly.)

Assuming the computations come to the same conclusion.

But e.g. 96-bit extended floating point calculations on Intel/AMD hardware will generally give different answers than 128-bit Sparc or Altivec calculations, and I'm sure you encounter similar problems in the high-end graphics market.

Compare: Matlab's Loss is Nobody's Gain or How JAVA's Floating-Point Hurts Everyone Everywhere . (Scroll down to the bottom of the page; they're PDF files, which is why I didn't link to them directly.)

This was my initial belief, but I looked up the definition online and the definition as used in biology was this: "Change in the genetic composition of a population during successive generations, as a result of natural selection acting on the genetic variation among individuals, and resulting in the development of new species." According to this definition, at least, you can't separate the theory of evolution from that of natural selection. Looking up "natural selection" seems to confirm this, "evolution" is just short for "Darwin's theory of evolution", and it seems to me that natural selection, as the sole basis for the origin of human life, is far from scientific fact.

There's a lot of misinformation out there, and you're going to find a lot of it online. The fact of the matter is, evolution can't possibly be synonymous with Darwin's theory of natural selection, because scientists had been actively trying to figure out how evolution works long before Darwin was on the scene. Look up Lamarckian Evolution to read up on an older competing theory for how evolution works. It was still fairly popular as recently as 100 years ago.

In fact, there is some more recent research showing that there might be Lamarckian elements to evolution, which would necessitate the replacement of Darwin's theory of evolution by natural selection with a new one. Let's sincerely hope that the law and the theory are separate entities, because otherwise we're going to have to come up with a new word.

Oh, and here're some more:

http://my.execpc.com/4A/B7/helberg/pitfalls/
http://www.stat.berkeley.edu/users/stark/SticiGui/ Text/ch16.htm

With information like this available about the misuse of statistics, I find the crap that comes out of the current Presidential administration amusing. Things like jiggling the numbers when reporting the number of wounded and dead from Iraq, employment numbers, Social Security liquidity, that sort of thing. Understanding how statistics are used as propoganda tools makes it easier to recognize a liar when you hear him give the State of the Union address.

And for some more black-hole info:

Black Hole FAQ

And on a side note, it's been a long time since I've watched my DVD of "The Black Hole", so I may have to do that now :). The last time I watched it, I was surprised how dark it was (no pun intended) for a "Disney Movie". May also have explained why I liked it so much as a kid...

N.

Doesn't correlate with here or here

There is some practical problems.

1. Low yield. Failure rate grows with bigger chips. Makes them more expensive.
2. Harder to make. Different prosesses for making memory and logic (this has been done already of course).
3. Heat problem. It's easier to cool separaate chunks of prosessor, graphics processor and memory.
4. Upgrading. New usb spec, make new mask. Upgrade graphics prosessor, make new mask. New mask for every memory configuration. Uh.

We may get close eventually. Practicality may dictate that we end up with 1-3 chips per home PC. Maybe optical connections between.

My time estimate for this to happen is 10-30 years from now.

Really Logo wasn't intended to teach programming (though of course it did that). It was intended to teach math, and algorithmic thinking, and thinking in general. And, paired with the right teacher and an interested pupil, it's really great at that. Without realizing it, a child can end up learning not just geometry (through the turtle graphics), but a lot of pre-algebra. I think programming is a far more accessible way to introduce algebra than the traditional techniques; even young children can understand variables in programs, when the declarative variables that are used in mathematics are much more challenging.

It's also a better language than many of "teaching" languages, like Basic. It's an old-school version of Lisp, with a little tweak to avoid the parenthesis. And don't be fooled by things that call themselves Logo when they are just turtle graphics. Turtle graphics are cool, but just a piece of the equation. (Though not-so-coincidentally, Python has built-in turtle graphics).

If you are really interested in programming as education, I might recommend the book Mindstorms, which is a classic about some of the theory behind teaching with Logo. It's not a practical guide, though many of those also exist.

If you are looking for a Logo implementation, on Windows I would recommend Elica, MSWLogo, and UCBLogo, in that order. On Mac or Linux, you can use UCBLogo, Turtle Tracks (a cross-platform Java implementation), or on Mac one of a number of (rather expensive) commercial Logos. If you are a programmer and feel like fiddling alongside your child, you might try my project PyLogo, which is cross-platform and written in Python, but quite rough around the edges. Or if you want something that is Logo, but pretends to be a general-purpose scripting language, look at Rebol. Or for a slightly-lame but functional embedded robot Logo, Cricket Logo. Or for older people, NetLogo is a massively-multitasking implementation to use to play around with autonomous entities (e.g., ant simulations). NetLogo is kind of the successor to StarLogo.

For more information on Logo, you can look at the Logo Foundation, or get in touch with many helpful users in the LogoForum Yahoo Group.

I think Logo is great for kids of that age because it provides them with instant gratification at the early stages. Once they get past seeing what the commands do at the Logo prompt, I have them write short programs using a text editor.

Some links about the site...

• Topographic Mapping at the SETI Radio Telescope Observatory (UC Berkeley)
• Topographic map of site (Topozone.com)
• "The big idea in SETI: Think small" (MSNBC, Feb 8, 2004)

See nucleosynthesis.

This is slightly unrelated.

The article is about SETI, not SETI@home.

The latter is simply part of the larger SETI project that involves several telescopes, and independent computers looking for signals full time. SETI@home is used for processing lower priority areas of the SETI data.

by the feds to catch those people laser-sniping the aircrafts.

I wasted this http://setiathome2.ssl.berkeley.edu/fcgi-bin/fcgi? email=webmaster%40x-term.com&cmd=user_stats_new many CPU-cycles for SETI@home.

the US millitary Already has a working (leg only) exosuit. Right now it's meant as a gear/supply transport platform... The current practical applications would allow a land based party of infantry to carry rations and munitions for miles and miles away without needing any airdrops of supplys or ther currently employed methods for getting goods into terain not accessable by conventional vehicles.

Some possible configurations for the suit on a 'real' battle field include a 'practical' flame thrower. you could load half a ton of propellant onto the back of this suit and have enough burntime to kill 100,000 infantry*.. or you could have a really big power plant on the back to power a tank piercing laser... and it could have low power for soldiers... or you could add a missle battery to it, or you could havs two really heavy gattling guns with about 1,000,000** rounds of ammo... Heck, while were at it let's make it a platform for your own personal Ion cannon! because you know it's not enough to just cut a hole in a tank, it needs to be entirely melted into a pile of volcanic debris!

*= well i'm guessing if they group up for you nicely...

**== well a lot anyways.

Consider joining a BOINC project. Runs on Windows, Linux, Mac, or anything you can compile on. The Slashdot team on SETI@home for BOINC has 556 members.

Consider joining a BOINC project. Runs on Windows, Linux, Mac, or anything you can compile on. The Slashdot team on SETI@home for BOINC has 556 members.

Download the BOINC/SETI client now:

http://setiweb.ssl.berkeley.edu

It also currently works on mac osx, linux, solaris.. with graphic versions of each of these being tested as I type. And it's open source, of course.

We just got a bunch of new hardware to handle the load as we transition the 500,000 active SETI@home classic users to BOINC. This will happen very early in 2005.

- Matt - BOINC/SETI@home

Download the BOINC/SETI client now:

http://setiweb.ssl.berkeley.edu

It also currently works on mac osx, linux, solaris.. with graphic versions of each of these being tested as I type. And it's open source, of course.

We just got a bunch of new hardware to handle the load as we transition the 500,000 active SETI@home classic users to BOINC. This will happen very early in 2005.

- Matt - BOINC/SETI@home

I really like Folding@Home too.

Folding@Home is pretty to watch as the atoms get twisted about. I particularly like when you get some of the larger projects with 12,000+ atoms in total.

I used to crunch SETI packets, but that project lost me after when they upgraded to BOINC. BOINC managed to crash my otherwise unshakable Win2K machine, which is completely unacceptable.

So here's my question: How pretty is the eye candy on this new IBM client? Does anyone have screenshots?

I really like Folding@Home too.

Folding@Home is pretty to watch as the atoms get twisted about. I particularly like when you get some of the larger projects with 12,000+ atoms in total.

I used to crunch SETI packets, but that project lost me after when they upgraded to BOINC. BOINC managed to crash my otherwise unshakable Win2K machine, which is completely unacceptable.

So here's my question: How pretty is the eye candy on this new IBM client? Does anyone have screenshots?

I just realized what I did. I get it. Pr0n. Boink. BOINC.

ba-dum dum. I was refering to the Berkeley Open Infrastructure for Network Computing (BOINC)

BS?

Same press release on a faster server.

kernel: The semantics of setuid class of calls is very confusing. Worse, it varies from from one flavour of Unix to another.
Look at David Wagner's "Setuid Demystified" paper The graphs in pages 9, 10 and 11 are scary.


user level programs:
I use Microsoft Windows sometimes and I find the "system restore" to be a nice feature. It is just convenient to press a button than worry about depdencies when you screwed up etc. Especially useful is you are sending a laptop to your computer illiterate mother in India and expect that she is probably going to try and install some random programs. If something goes wrong, I can ask her to turn the clock back, sitting in the other side of the world.



Sudhakar

> 1500 people trying to control one webcam simultaneously is a ridiculous thought.

Some links:

• ShareCam, ALPHA Lab, University of California, Berkeley.
• Info on observer.berkeley.edu and its FAQ.
• "The co-opticon (a.k.a. the ShareCam) is a machine for democratic optics, allowing a network of participants to cooperatively control the viewpoint of a shared video camera. The co-opticon combines a networked robotic video camera with a graphical user interface that allows many internet-based viewers to share simultaneous control of the camera by specifying desired viewing frames. Algorithms compute the optimal camera frame based on all requests, and position the camera accordingly."
• demonstrate.berkeley.edu. The installation combines the world's most advanced networked robotic camera, a visual database, and a mathematical model of socio-ocular camera control.

-j.

P.S. Coral cache acting weird from here. Couldn't coralize all the links.

> 1500 people trying to control one webcam simultaneously is a ridiculous thought.

Some links:

• ShareCam, ALPHA Lab, University of California, Berkeley.
• Info on observer.berkeley.edu and its FAQ.
• "The co-opticon (a.k.a. the ShareCam) is a machine for democratic optics, allowing a network of participants to cooperatively control the viewpoint of a shared video camera. The co-opticon combines a networked robotic video camera with a graphical user interface that allows many internet-based viewers to share simultaneous control of the camera by specifying desired viewing frames. Algorithms compute the optimal camera frame based on all requests, and position the camera accordingly."
• demonstrate.berkeley.edu. The installation combines the world's most advanced networked robotic camera, a visual database, and a mathematical model of socio-ocular camera control.

-j.

P.S. Coral cache acting weird from here. Couldn't coralize all the links.

• Firefox
• Thunderbird
• Outpost Firewall
• Cygwin
• The GIMP
• Spybot
• adAware
• Trillian
• Google Desktop Search
• SETI@home
• iTunes

My former advisor here at UC Berkeley, Gibor Basri, has a neat way of discriminating between planets and the lesser (comets, asteroids, etc.). His idea is that if the object has enough self-gravity to force it into a spherical shape, it's a planet... if it doesn't (like Mars' "moons"), it's something less.

Here's a snipet:

How can this be resolved? A consensus is slowly developing (I believe) for the following solution. We can first define what we mean by "planetary mass", and base this only on physical characteristics. Then we can include circumstance into the definition of "planet". I propose the following three definitions:

FUSOR - an object that achieves core fusion during its lifetime.

PLANEMO - a round non-fusor.

PLANET - a planemo orbiting a fusor.

[...]

read on for his full article [berkeley.edu].

The following is a draft of an article now published in the Nov/Dec 2003 issue of Mercury. Draft of Mar. 20, 2003.

Defining "Planet" by Gibor Basri Univ. of California, Berkeley

Even before they were civilized, people looked into the sky and recognized different celestial objects. The Sun defined daytime, and the stars provided a fixed background of faint, twinkling lights at night. Among them moved the Moon, and a few special steadier lights. The Greeks called those which moved "planets" (it is worth noting that the Sun and Moon were originally included, since motion against the stars was the defining characteristic). Most cultures have an analogous word for these "wanderers". Both the stars and the planets were thought to revolve around the Earth.

After the Copernican Revolution, we recognize the Moon as the only body that orbits the Earth. The Sun is a very nearby example of a star, and the visible planets are other large bodies that orbit the Sun. We see them by reflected sunlight, while stars produce their own visible light. This understanding yields the dictionary (lay public) definition of the word "planet": a large heavenly body that shines by reflected light and orbits the Sun. In the past century we gained much understanding of our Solar System, and even visited most of the planets robotically. Yet today, professional astronomers find themselves unable to agree upon a succinct definition of "planet". Replacing "the Sun" with "a star" is obviously necessary now that many extrasolar planets have been discovered, but the problem goes well beyond that.

Two recent controversies that found their way to the popular press illustrate further difficulties. One is the "Pluto controversy". This arose because of the discovery of a large belt of icy objects beyond Neptune. They are the outer remains of the original protoplanetary disk. This "Kuiper Belt" is a natural outcome of incomplete planet formation in the outer Solar System, and is the source of some of the comets we see. As Kuiper Belt objects (KBOs) were discovered in increasing numbers in the 1990s, including a population of "Plutinos" which share Pluto's orbital characteristics (somewhat different from the other planets), some astronomers began to suggest that Pluto itself (which shares many properties with, but is the largest KBO known so far) does not qualify as a planet. The recent discoveries of Varuna and Quaoar (which are KBOs half the size of Pluto, like its moon Charon) may presage the time when we find another Pluto-sized KBO.

The current situation is much like that in the early 1800s, when the first asteroids were discovered. Ceres was originally hailed as the fifth planet, particularly since one in its position was expected from "Bode's Law" of planetary spacings. It lost its status within a few years, when other members of the asteroid belt began turning up. Herschel, who had been the only person to have discovered a new planet before then, aided the effort to demote Ceres. The arguments

Of course, we're constantly bombed with xrays from outer space too, though.

does not prove that diebold's systems systematically and intentionally pushed the election in any particular direction

Every Home a Prison by Jello Biafra.

But the point of the quote is of course still appropriate, as anybody who's worked with redundancy in any field would be aware.

Actually the functionality that you're looking for is available, albeit in protoypes in academia. Marc Davis at UC Berkeley (formerly of MIT Media Lab) has been working on automatically capturing video metadata and storing using the facilities of MPEG4 in conjunction with sophisticated searching. Have a look here.

• a PhD student or his advisor has an idea, writes papers about it and maybe even implements it. Then he starts a company and they make milions with it. Synopsys and Cadence (the two biggest players of the market) were started like that.
• The same people have an idea and actually have a full implementation and they sustain the basic research. The tool is put in open source fashion (rarely is GPL'd) and every company can modify it. The shiniest example is SPICE . The first version was written more than 30 years ago. Berkeley still owns it and everybody implements variants that are more or less compatible. Either Commercial eda tools (HSPICE, spectre, eldo, adssim) or proprietary implementations that are used within a company (TI spice, ST spice, motorola own spice, etc.). Analog design wouldn't exist without spice.
• Companies try to dominate the market. They figure out that they need to develop and control a platform. They make it open source (of course all its products work on it, more or less). An example is Cadence with open access. The idea of course is opposed by competitors who try to pass their own platform. Eventually they will reach an agreement

• Berkeley (Bancroft Library)
• UCLA (Young Research Library Special Collections
• http://orpheus.ucsd.edu/speccoll/duppol.html

A link which backs the "greedy bastards" theory :
http://math.berkeley.edu/~kirby/journals.html

It is being actively developed. While William Schelter was maintaining it (for 19 lonely years), development was very slow indeed. I gather that most of the work was done by him, and some of his graduate students. Since his death in 2001, a number of other people have come on board, and there is a lot of catching up to do.

Some documentation has been rewritten, a great many bugs have been squashed, the package has been ported to several Lisps (yes, it does matter to users), there has been at least one new Emacs mode written for it, it can be used with Texmacs, and so on. Some of the people who are working on it are big names in their spheres, like Richard Fateman, who worked on the original Macsyma.

Version 5.9.1 was released in September '04, and the next big step will be the GREAT SOURCE DOWNCASING. Maxima is so old that most of it is written in all caps. There is a lot to do to bring it into the 21st century, and most of what's being done right now is behind-the-scenes stuff.

As you say, it's decent software now. It's fully usable, with a useful GUI for Windows (developed on Schelter's watch, as I recall). It is probably better for memory intensive work than Maple or Mathematica; that's what initially got me started using it.

limited life (10% efficiency in 10 years for most designs)

As for your comparison to other semiconductor products, you are confused. Solar cells do not require multiple etching steps; you can grow a doped crystal, cut it into pieces (and save the dust for reprocessing), dope one side of it the other way to create the junction, and add contacts. You're not masking, cutting trenches, sputtering multiple layers of interconnects or any of the chemical-intensive processes which characterize LSI chips, so any claim of pollution which draws on LSI production data is inherently flawed.

Skip Myhrvold (wtf?) and put in William Kahan (only the driver of IEEE 754). This list is about comp. sci., not get-rich-quick schemes.

He was using Berkeley's computer system, so I doubt that Eolas' allegation of "not being connected to the internet" can be considered valid.

In addition to the system he used to create Viola, there's the copyright date on this article (1994), and this review of ViolaWWW by Tim Berners Lee (1992). Although there is no indication of whether Pei's embedded object archetecture predate's Eolas' the publication of his article and the availability of the libraries and samples at the same time that Eolas is patenting their new technology does cast doubt on Eolas' claim.

Eolas' argument is crap, they were likely aware of Tim Berners Lee's article and of Viola when they developed their techjnology (if you are an early developer of a new technology, it is likely that you read the works of the man who invented it), and the judge either doesn't understand the issue or the history, or he has an axe to grind (either against Microsoft, or against things being "Free").

A little narrative to go with your pictures -- check out this gripping account of this very earthquake by author Jack London (of Call Of The Wild fame).

Aww, come one. Kite aerial photography. There's been links about that on Slashdot before. That link talks about events which happened 98 years ago!

I know Slashdot sometimes report on old news or dupe, but that's ridiculous.


;)

What a bunch of nerds!

Do you really trust these dorks to predict your earthquakes? C'mon, you got to put some polish on this research if you want it to sell, baby!

The ViolaWWW browser application has provisions for treating viola application files just as any WWW document-- transport via HTTP, and render mini viola applications as if they're any web data.

Has written a comment to this case: http://www.xcf.berkeley.edu/~wei/viola/aboutEolasM icrosoft.html

http://www.xcf.berkeley.edu/~wei/viola/aboutEolasM icrosoft.html

Discussions about paper trails, auditing and such are for future elections. Only votes (not necessarily ballots) which have an origin which can be proven should be accepted. All others should be discarded. Without being able to prove the provenance of the "votes" it cannot proven that there was no monkey business. That's a lot of what having an open and fair election is about - proving there was no monkey business.

There are visible discrepancies between exit polls and poll results indicate that something is amiss and should be investigated.

Among several probelms, there seems to be a correlation between use of voting machines and skewed poll results as show by several investigations, including the Berkeley report. (Of interest for future reference, other analysis suggests a correlation between red counties and ClearChannel market dominance). Furthermore, the last few years of use have shown that the technology and methods used by the electronic voting machine companies that this little or no chance of preventing tampering nor of directly detecting tampering due to inherently badly designed systems. This was known prior to the election.

Such serious things should not be given the benefit of the doubt. The electoral college should overturn these irreproducable results and hold a second, paper-only election. It would have been both faster and cheaper to count them by hand like in Canada. Don't let it get so bad that it gets resolved by another civil war.

Personally I think the money they spent on the iraq war would have been better used at NASA. But your right, the US may find itself in deep financial trouble in the future....

Personally, I'm against war. Personally, I know Bush W is an idiot and a liar.

However, after 3 years of looking, I found out why we went to war in Iraq, and as much as I am against war and Bush, I believe that killing about 16,000 to 20,000 people in preservation of the US economic status is completely worthwhile.

Sources:

http://ist-socrates.berkeley.edu/~pdscott/iraq.htm l

http://www.feasta.org/documents/papers/oil1.htm

BitTorrent trackers just give you a way of finding peers who are downloading the same file - they are *not* necessary for data verification. A P2P search network like CRL would allow you to find peers that are interested in the same (verifiable) filename in a completely decentralized way. You could then use BitTorrent's parallel, incentive-based download mechanism to retrieve the file.