Slashdot Mirror

I have found that Code Crusader is a good IDE. It's easy to use and GPL'd.

I was REALLY disappointed in what I saw and read about Jesse! This CAN'T be what SGI develops in, can it?

Quicker

Try:

Avida

Tierra

I submitted this a few days ago, with a couple interesting links:

Here is the researchers web site.

and Here is a page related to the Nature article which includes source code for the experiments. They ran this on a Linux Beowolf cluster.

Project Von Neumann is apparently some effort to create and artificial life game. Dunno much about it, doesn't look like much coding has been done. Take a look for yourself.

You can find pictures of the sites somewhere off the LIGO page.

Check this out: ;-)

It is my understanding that a faulty computer chip had a bug in it that caused a cover over the dewer to be released too early during launch. This unfortunately, heated the dewer and the hydrogen cryogen then vented at a relatively extreme rate. Since the cryogen is completely vented, the detector cannot be cooled. The detector must be cooled in order to measure the low amounts of radiation that it is intended to measure.

I don't know if their web page is official or not, but it has many pictures. Actually, I could have saved the typing as the first entry (1999 March 29) describes the failure. So read it there.

~afniv
"Man könnte froh sein, wenn die Luft so rein wäre wie das Bier"

Not yet, but maybe you can add it for him :).

My dad teaches at Caltech, and published this paper on his website.

Various other people pointed to broken protocols, and protocols which need special help. In general, any protocol which does not restrict itself to a single connection (ie. src ip/port dst ip/port quad) will require special assistance. This includes FTP (both passive and active) in the general case, although for simple masquerading passive ftp does not need help.

For static NAT, where an IP address is always mapped the same way (n:n NAT, eg. 192.168.1.* is mapped straight into 1.2.3.*), only protocols which actually include IP addresses within their data stream will be impaired. Unfortunately, FTP is one of these.

A special note on games: Dan Kegel (of Activision) produced a fairly well-thought-out proposal for UDP gaming through NAT. IP masquerading in Linux 2.2 meets this standard.
Here is the draft

Rusty.

I've played a fair bit of games from my Win98 box behind the IP Masq, and for many newer games they work just fine (playing, not hosting).

Those that don't need to get their act together. :) This has some explanations of a method to use UDP packets and work beautifully with different NAT systems.

Things that I have played just fine recently (read, I at least see their CDs lieing around my desk, more work but I can't think of em all right now): Half-Life, Quake 3, Myth 1 and 2, Tribes, F22 Lightnine 3 Demo, even 2am.com's group of free games. I did pop in my old SWAT 2 and that one didn't work. Some game companies at least have a tech support FAQ that may tell what ports to redirect or anything to help. I say we start petitioning companies that refuse to make Linux ports to at least make compatible multiplayer gaming...

Let's have a show of hands: how many of you have actually seen a Beowulf cluster? Where I work (Caltech's CACR), we have a 114-node system, and it's pretty damn big. You can't just have all of the nodes packed densly: you need to be able to access the backs for networking, power, etc. It takes up a pretty sizeable amount of floor, and reaches up to the (rather high) ceiling.

Here's a couple of pictures. The one up top is just one side of it.

These guys want to make one that's over 100 times as big! Can you imagine the network cable nightmare? Not to mention the power requirements. Makes you feel sorry for the technician that has to set it up.

The other big problem with a large cluster is network latency. You can reduce the effects of this by passing larger packets of info, but there's still a limit that you reach. Just because you make something 100 times bigger doesn't mean it'll be 100 times better.

I also think that the software configuration would play a major role in the efficiency. I'd rather trust trained scientist (not me; I'm just a student), who's been working with large-scale parallel machines for years to set this up, not some tech guys who thought it'd be a neat idea. But maybe I'm just pessimistic.

Still perfectly happy with my 1-node PII... -ElJefe

Let's have a show of hands: how many of you have actually seen a Beowulf cluster? Where I work (Caltech's CACR), we have a 114-node system, and it's pretty damn big. You can't just have all of the nodes packed densly: you need to be able to access the backs for networking, power, etc. It takes up a pretty sizeable amount of floor, and reaches up to the (rather high) ceiling.

Here's a couple of pictures. The one up top is just one side of it.

These guys want to make one that's over 100 times as big! Can you imagine the network cable nightmare? Not to mention the power requirements. Makes you feel sorry for the technician that has to set it up.

The other big problem with a large cluster is network latency. You can reduce the effects of this by passing larger packets of info, but there's still a limit that you reach. Just because you make something 100 times bigger doesn't mean it'll be 100 times better.

I also think that the software configuration would play a major role in the efficiency. I'd rather trust trained scientist (not me; I'm just a student), who's been working with large-scale parallel machines for years to set this up, not some tech guys who thought it'd be a neat idea. But maybe I'm just pessimistic.

Still perfectly happy with my 1-node PII... -ElJefe

I can vouch for the cool-ness of this work. I'm an undergraduate at Caltech doing research in Christof Koch's lab, one of the big participators in Telluride. If you want more information, you can visit the Koch Lab homepage .
While I'm only doing psychophysics research right now (just finished up my freshman year, so the mathematics isn't quite up to snuff yet), I'm wondering what amazing things I'll be studying once I start working on my Ph.D. thesis :-)

Nick Knouf
nknouf@cedric.caltech.edu
nknouf@klab.caltech.edu
http://cedric.caltech.edu

I can vouch for the cool-ness of this work. I'm an undergraduate at Caltech doing research in Christof Koch's lab, one of the big participators in Telluride. If you want more information, you can visit the Koch Lab homepage .
While I'm only doing psychophysics research right now (just finished up my freshman year, so the mathematics isn't quite up to snuff yet), I'm wondering what amazing things I'll be studying once I start working on my Ph.D. thesis :-)

Nick Knouf
nknouf@cedric.caltech.edu
nknouf@klab.caltech.edu
http://cedric.caltech.edu

Ditto to that.
about Carl
Carl J Lydick photos

Yes, there are at least two:

Honda made a pair of bipedal robots, the P2 and P3, that ar capable of walking, including up and down stairs, and pushing a cart around. The Robot page is in Japanese, so I can't tell too much more about it. (Sorry, my Japanese is REALLY bad.) Watching it walk is kind of eerie, it looks just like a human walking, just slightly hunched over. They both can walk 2 km/h (about 1.25 mph.)

The P2 is 1820mm tall (5'11.6") and weighs 210 kg (463 lbs), and the P3 is 1600mm tall (5'3") and weighs 130 kg (286.5 lbs.)

The P2 has a large, boxy head and a huge, boxy "backpack", whereas the P3 has a round head and a much smaller "backpack".

There is a page with a bunch of video clips (2 RealVideo, 5 embedded QuickTimes.) Warning, it's a BIG page for modem users. (Hooray for ADSL!)

Also see this page for information on it from a researcher at Cal Tech.

Oops... After writing all that, I noticed an English version of the robot info page... Sorry, I'm too lazy to go back and edit all those URLS.

code crusader's pretty nice..

www.cco.caltech.edu/~jafl/jcc/

See also:
http://www.newscientist.com/ns/9 81107/nshorts.html for a short note about the paper published in Science (Search for Benzer, free registration needed to read the abstract).

There is also a fantastic article on Seymour Benzer in the April 5 issue of the New Yorker.

I think that you meant 331372 bytes and 3296 bytes.

Take a look at

http://www.cco.caltech.edu/~ja fl/jx/egcs_complain.html

Generally, most of the data volume seems to come from reading out large CCDs. For example, a friend here at Caltech is doing fluid dynamics research in which they read out a 1024x1024 array at 1000 fps for a couple of seconds (I think the pixels are 12 bits each or so). This isn't too different from the interferometers, which are creating data at sustained rates of around 5-10 KHz. At those volumes, it doesn't take long to fill a terabyte, and when you go to long-term storage, you're into petabytes. Of course, for time-series image data like this, lossless differential compression schemes can be a big help (by a factor of four or ten), enough to save a few million dollars of storage space.

Generally, most of the data volume seems to come from reading out large CCDs. For example, a friend here at Caltech is doing fluid dynamics research in which they read out a 1024x1024 array at 1000 fps for a couple of seconds (I think the pixels are 12 bits each or so). This isn't too different from the interferometers, which are creating data at sustained rates of around 5-10 KHz. At those volumes, it doesn't take long to fill a terabyte, and when you go to long-term storage, you're into petabytes. Of course, for time-series image data like this, lossless differential compression schemes can be a big help (by a factor of four or ten), enough to save a few million dollars of storage space.

I think that the theory to which they are referring to is the thought that large gas giants (like Jupiter) tend to form at large distances than solid, rocky planets (like Earth). Two of the discovered planets are closer to the sun than Earth is.

Also, large planets close together tend to disturb and destabilize each others orbit. Three-body systems are very unstable; four bodies even more so.

More info on 3-body problems:

FYI:
gcc and egcs are available for download from SCO. It's true that many programs require some porting to compile under the SCO supplied complier, but normally it's not difficult.

There is also lxrun which lets you run many linux binaries without recompiling.

Many of the tools you "go looking for" are also probably at the SCO Skunkware site for free download.

Thus Code Crusader still can't go in Debian proper. (Though in Debian's classification it is now suitable for "contrib", whereas it was only suitable for "non-free").

Thus Code Crusader still can't go in Debian proper. (Though in Debian's classification it is now suitable for "contrib", whereas it was only suitable for "non-free").

MSWLogo is an extension of UCBLogo that includes a lot of more graphical things, but it only works on Windows :(

I also recently saw a bit about a Java version of Logo. You can find more about that at here.

I'd highly encourage teaching Logo to a child. It's a great language, far far better than Basic.