I just installed Fedora Core 2 on a P-II 233 with 64 megs of ram (aka "Computerstein", the computer built from spare parts taken out of my good computer). First few times I ran the installer, it spontaneously rebooted in the middle of install. I increased the swap partition size and installed in text mode and it worked.
It can't start up the default Gnome environment without going heavily into swap. Incidently, does anyone know how to turn off nautilus? (And no, kill doesn't work -- it respawns itself. Argh!) I know how to turn it off on Mandrake, but the relevant menu item doesn't seem to be present in Fedora.
The single-xterm-only mode of starting X runs tolerably fast.
For those who aren't familiar with citeseer, it's a publicly-accessible database of scientific literature. There's a downloadable pdf available for most papers.
The entry for every paper has links to papers it cites, links to similar papers (i.e. papers that cited this paper were also likely to cite these papers), and a citation count (which can be a good way to estimate the relative importance of a given paper - if something has been cited 300 times, its probably worth a read).
Looks like frys is selling an xp2500+ with motherboard for $65 US (If you happen to live in the Dallas area which I don't). Is there a market for processors cheaper than that?
Am I the only one who thinks of celery every time I hear the name "celeron"? I don't have such an immediate word association with sempron, so I guess that's a good thing.
[you use radiosity] to compute global illumination levels ahead of time (as opposed to using ambience). Then you can use raytracing in realtime to capture dynamic effects.
Where's the fun in that? I'd like to see dynamic global illumination (It might take awhile before computers are fast enough, and people figure out the right algorithms, though).
Photon mapping can be precomputed as well, and can simulate a number of effects that can't be simulated with radiosity (subsurface scattering, caustics off of shiny objects).
I was under the impression that ray tracing and radiosity weren't exclusive techniques, and using both produced excellent results.
Yes, radiosity and photon mapping are both usually used to calculate indirect lighting. Usually they're used in conjunction with ray tracing or polygonal rendering.
Id argue that the most valuable commodity Nintendo has at the moment is the Pokemon franchise.
(shudder) While you might be right in terms of current cash flow, Zelda has the appeal of a wider age group, since many of us grew up on Zelda games. Zelda thus has greater long-term value.
Radiosity would dramatically increase the computational complexity.
Polygonal rendering: O(N), where n=number of triangles
Ray tracing: O(log N), where n=number of objects (assuming a good bounding volume heirarchy)
Photon mapping: O(P log max(P, N)), where P=number of photons, which generally must be inserted into a kd-tree, and N=number of objects
Radiosity: O(N^2), where n=number of triangles
Ray tracing could conceivably make a game faster, if the scenes are complicated enough. Radiosity, on the other hand, is very very slow. Photon mapping might be a better choice - it traces rays from the light source, and stores photons at the object intersection points, which are then used by the ray tracing step to approximate global illumination.
thus fulfilling its original purpose: write once, run anywhere.
It already does that. The matter of open source is irrelevant.
It's not "run anywhere" if a large portion of the user base has good reasons not to install the JVM (for instance, its not worth the trouble of manually installing, and I don't like having more untrusted binaries on my system than I have to).
That thing's almost as fast as a hard drive. I wonder if it would be possible to use it to store the journal file for ext3, to avoid writing data out to the hard drive twice. Or maybe it could be used to hold the swap partition (no seek time == fast page faults), though buying more ram would no doubt be cheaper.
I've played quite a bit of A&A myself, though these days I prefer Shogun (aka Samurai Swords). There's less tedious piece-moving, and some of the less exciting steps (like buying units) can be performed in parallel. I also like how the turn order isn't fixed, the starting conditions are randomized (much like risk), and alliances aren't fixed.
That's a reasonable strategy until the other program overwrites your tight loop with an illegal instruction (like DAT) and all your threads die when they execute it. There's no memory protection in core wars, that's what makes it interesting. Its often necessary for your program to make copies of itself in memory as well as fork additional execution contexts. Also, your threads run slower than your opponents if you have more of them running. (At least, this is my basic understanding of the game. I never really spent enough time to fully understand it.)
Pardon, but I can no longer contain my pedantry... I trust to be forgiven, since there is a fairly precise definition of the term you are bandying about.
I can certainly forgive your pedantry if you can forgive my ambiguity.
If you have loops you don't need "if", by the way. And vice versa.
While for the most part I agree with you, if/then/else cannot simulate loops because it provides no way to execute a block of code more than once, and no way to remember state (like how many times we've executed the loop). If statements in conjunction with gotos and variable assignments, on the other hand, can be used to simulate a for loop with a counter, or just about anything else.
Just to make sure the title of this post has some relevance to its contents, I'd like to remind everyone that html has none of these features (unless you use javascript).
Tables can be pretty involved, if you a just using notepad as an editor.
Just because something is hard doesn't make it a programming language. There's a certain number of minimum features a language must support before its considered turing complete. (These usually include things like variable assignments, if statements, and loops.) HTML exposes people to the tedium of programming without giving them any of the real power that programming languages can provide. Javascript is another story, but as I haven't programmed in it, I really can't comment on its suitability as a first language.
I wonder if many people have used icws94 as a first language?
(For those that have never heard of core wars: the basic idea is you write assembly programs that run in a virtual machine - whichever program has more threads running at the end of a time limit wins. I never got into it, but it looks like fun.)
Have you heard The Ballad of Optimus Prime? (the link is just a clip, but I heard the whole thing a few weeks ago -- the funniest song I've heard on the radio in quite some time)
The SAN device looks cool - it appears to be a device that makes any USB mass storage device appear as a network drive. I wonder if it shows up as a windows file server, nfs server, or a network attached block device? Can two computers use it simultaneously? Does anyone else make something similar?
Not at all true! AMD64 has twice the number of general-purpose registers available in 64-bit mode. Some apps also just run faster in 64-bit, like POVray.
Don't forget the new nonexecutable page flag! It reduces the severity of buffer overflows, which is far from useless in my opinion. (I'm not sure if the processor has to run in 64 bit for it to work, but as far as I know, its only available on the 64 bit chips -- someone correct me if I'm wrong.)
My understanding is that an access point tells associated devices to transmit one at a time, in an ad-hoc network devices can transmitt over each other.
In PCF (point coordinated function) mode, that would be true - the access point polls each client in a round robin fashion, asking if they have data to send. However, PCF is an optional part of the 802.11 spec that has never been implemented.
In DCF (distributed coordinated function) mode, which all 802.11 hardware uses, everyone transmits whenever they feel like, as long as no one else is already currently transmitting, and no CTS (clear to send) transactions have been heard recently. All transmissions must either be from, or to, the AP, though, in infrastructure mode. Sending a packet from client A to client B would require two hops -- one from A to the access point, and one from the access point to B. In an ad-hoc network, anyone can talk to anyone.
Vendor A could specify a signal of -80 dBm for a BER of 10**-6, while vendor B could specify -90dBm for a BER of 10**-3.
I think the standard practice is to quote the receive sensitivity for an 8% packet error rate. (Note: a single bit error will cause the whole packet to be discarded). I assume they use full size packets.
Independent testing would be nice, but I'm not sure any of the ususal sites (tom's hardware etc) have the motivation, knowledge, and equipment to meaningfully compare radio performance aside from simple range tests.
and RPM? well it does that... and it does it all better. I have yet to see a better packager than URPMI... ever.
I won't say it's better, because I haven't used it much, but fedora's yum seems quite similar to urpmi, and is easy to set up to get stuff from local mirrors.
Me, of course, I experiment, so I literally install EVERYTHING, and it still works!
I installed Mandrake today and was disappointed by the lack of a simple "install everything" option. What's the point of all those CDs if most stuff has to be manually selected?
Another minor gripe: it asks for the CDs again after installation. Why aren't all the packages for network configuration installed by default? Don't almost all users connect their boxes to the network these days? Argh!
(Note: despite my whining, I really do like Mandrake, I just felt obligated to add some diversity to the hundreds of "yay Mandrake" posts)
Slashdot Mirror
I just installed Fedora Core 2 on a P-II 233 with 64 megs of ram (aka "Computerstein", the computer built from spare parts taken out of my good computer). First few times I ran the installer, it spontaneously rebooted in the middle of install. I increased the swap partition size and installed in text mode and it worked.
It can't start up the default Gnome environment without going heavily into swap. Incidently, does anyone know how to turn off nautilus? (And no, kill doesn't work -- it respawns itself. Argh!) I know how to turn it off on Mandrake, but the relevant menu item doesn't seem to be present in Fedora.
The single-xterm-only mode of starting X runs tolerably fast.
-jim
For those who aren't familiar with citeseer, it's a publicly-accessible database of scientific literature. There's a downloadable pdf available for most papers.
The entry for every paper has links to papers it cites, links to similar papers (i.e. papers that cited this paper were also likely to cite these papers), and a citation count (which can be a good way to estimate the relative importance of a given paper - if something has been cited 300 times, its probably worth a read).
-jim
Looks like frys is selling an xp2500+ with motherboard for $65 US (If you happen to live in the Dallas area which I don't). Is there a market for processors cheaper than that?
-jim
Am I the only one who thinks of celery every time I hear the name "celeron"? I don't have such an immediate word association with sempron, so I guess that's a good thing.
-jim
Where's the fun in that? I'd like to see dynamic global illumination (It might take awhile before computers are fast enough, and people figure out the right algorithms, though).
Photon mapping can be precomputed as well, and can simulate a number of effects that can't be simulated with radiosity (subsurface scattering, caustics off of shiny objects).
-jim
Yes, radiosity and photon mapping are both usually used to calculate indirect lighting. Usually they're used in conjunction with ray tracing or polygonal rendering.
-jim
(shudder) While you might be right in terms of current cash flow, Zelda has the appeal of a wider age group, since many of us grew up on Zelda games. Zelda thus has greater long-term value.
-jim
Radiosity would dramatically increase the computational complexity.
Polygonal rendering: O(N), where n=number of triangles
Ray tracing: O(log N), where n=number of objects (assuming a good bounding volume heirarchy)
Photon mapping: O(P log max(P, N)), where P=number of photons, which generally must be inserted into a kd-tree, and N=number of objects
Radiosity: O(N^2), where n=number of triangles
Ray tracing could conceivably make a game faster, if the scenes are complicated enough. Radiosity, on the other hand, is very very slow. Photon mapping might be a better choice - it traces rays from the light source, and stores photons at the object intersection points, which are then used by the ray tracing step to approximate global illumination.
-jim
It's not "run anywhere" if a large portion of the user base has good reasons not to install the JVM (for instance, its not worth the trouble of manually installing, and I don't like having more untrusted binaries on my system than I have to).
-jim
That thing's almost as fast as a hard drive. I wonder if it would be possible to use it to store the journal file for ext3, to avoid writing data out to the hard drive twice. Or maybe it could be used to hold the swap partition (no seek time == fast page faults), though buying more ram would no doubt be cheaper.
-jim
It will certainly increase its adoption, especially in the open source world, thus fulfilling its original purpose: write once, run anywhere.
-jim
I've played quite a bit of A&A myself, though these days I prefer Shogun (aka Samurai Swords). There's less tedious piece-moving, and some of the less exciting steps (like buying units) can be performed in parallel. I also like how the turn order isn't fixed, the starting conditions are randomized (much like risk), and alliances aren't fixed.
-jim
That's a reasonable strategy until the other program overwrites your tight loop with an illegal instruction (like DAT) and all your threads die when they execute it. There's no memory protection in core wars, that's what makes it interesting. Its often necessary for your program to make copies of itself in memory as well as fork additional execution contexts. Also, your threads run slower than your opponents if you have more of them running. (At least, this is my basic understanding of the game. I never really spent enough time to fully understand it.)
-jim
I can certainly forgive your pedantry if you can forgive my ambiguity.
While for the most part I agree with you, if/then/else cannot simulate loops because it provides no way to execute a block of code more than once, and no way to remember state (like how many times we've executed the loop). If statements in conjunction with gotos and variable assignments, on the other hand, can be used to simulate a for loop with a counter, or just about anything else.
Just to make sure the title of this post has some relevance to its contents, I'd like to remind everyone that html has none of these features (unless you use javascript).
-jim
I was refering to this rather confusing slide, but lower frequencies do sound more plausible (not that I'm a radio engineer or anything)
-jim
Just because something is hard doesn't make it a programming language. There's a certain number of minimum features a language must support before its considered turing complete. (These usually include things like variable assignments, if statements, and loops.) HTML exposes people to the tedium of programming without giving them any of the real power that programming languages can provide. Javascript is another story, but as I haven't programmed in it, I really can't comment on its suitability as a first language.
-jim
I wonder if many people have used icws94 as a first language?
(For those that have never heard of core wars: the basic idea is you write assembly programs that run in a virtual machine - whichever program has more threads running at the end of a time limit wins. I never got into it, but it looks like fun.)
-jim
I'm surprised they got a radio link to work through 60m of ice. They're apparently using 1.8 ghz radios.
-jim
-jim
The SAN device looks cool - it appears to be a device that makes any USB mass storage device appear as a network drive. I wonder if it shows up as a windows file server, nfs server, or a network attached block device? Can two computers use it simultaneously? Does anyone else make something similar?
-jim
for something to support AODV or DSR out of the box... (might be awhile)
-jim
Don't forget the new nonexecutable page flag! It reduces the severity of buffer overflows, which is far from useless in my opinion. (I'm not sure if the processor has to run in 64 bit for it to work, but as far as I know, its only available on the 64 bit chips -- someone correct me if I'm wrong.)
-jim
In PCF (point coordinated function) mode, that would be true - the access point polls each client in a round robin fashion, asking if they have data to send. However, PCF is an optional part of the 802.11 spec that has never been implemented.
In DCF (distributed coordinated function) mode, which all 802.11 hardware uses, everyone transmits whenever they feel like, as long as no one else is already currently transmitting, and no CTS (clear to send) transactions have been heard recently. All transmissions must either be from, or to, the AP, though, in infrastructure mode. Sending a packet from client A to client B would require two hops -- one from A to the access point, and one from the access point to B. In an ad-hoc network, anyone can talk to anyone.
-jim
I think the standard practice is to quote the receive sensitivity for an 8% packet error rate. (Note: a single bit error will cause the whole packet to be discarded). I assume they use full size packets.
Independent testing would be nice, but I'm not sure any of the ususal sites (tom's hardware etc) have the motivation, knowledge, and equipment to meaningfully compare radio performance aside from simple range tests.
-jim
I won't say it's better, because I haven't used it much, but fedora's yum seems quite similar to urpmi, and is easy to set up to get stuff from local mirrors.
I installed Mandrake today and was disappointed by the lack of a simple "install everything" option. What's the point of all those CDs if most stuff has to be manually selected?
Another minor gripe: it asks for the CDs again after installation. Why aren't all the packages for network configuration installed by default? Don't almost all users connect their boxes to the network these days? Argh!
(Note: despite my whining, I really do like Mandrake, I just felt obligated to add some diversity to the hundreds of "yay Mandrake" posts)
-jim