SGI markets 1024 machines "out of the box" and there is a 6000+ CPU machine on the top 10. They are currently working on the 2nd generation which will allow 33,000 CPUs (15 bits of CPU's). It'll be cool.
My mistake re. SGI; thanks for making me aware of this (another poster pointed it out too).
Also, WRT address space, I would think memory access on these things is quite heavily abstracted for any userland tasks. When you reach outside of any one machine on the cluster, conventional memory access methods probably go out the window anyway. The ASCI Red was just a bunch of P6's soldered together after all, and it doesn't seem to be having too many problems.
I seriously doubt that any x86 cluster uses a unified address space from any given task's point of view.
Abstracting memory accesses could let you make your real address space a window into a larger one, but that would have some pretty nasty overhead.
They're awfully confident of McKinley not following in the footsteps of Merced if they've placed this order.
This raises an interesting question, though. If you want to build a high-performance compute cluster nowadays... what do you build it out of? The old answer, Alpha, doesn't really apply any more.
Sun is optimized for communications bandwidth, not FLOPS, and I'm not sure if SGI even _offers_ machines that huge. HP is betting on IA64. And x86 is competely unsuitable, for memory space reasons if nothing else.
It's easier to make a funny thing with a cheap Motorola 6800 or a Zilog Z80 than with a Intel586 or AMD K7. Both for the hardware side (it's only 40 pins and 2MHz) as for the software side (just a couple of registers).
You can buy a host of programmable microcontrollers from a variety of vendors; check Digikey's catalog for a sampling. Many of these should adequately substitute for a Z80.
Also, how "easy" is it these days to add an self-developped extensionboard into your computer?
Not that hard.
I was building a project driven off of a parallel port a couple of weeks back. These won't go away for a few years yet, and you can clock them as slowly as you want to.
You can also still find motherboards with ISA slots for new machines; at 8 MHz or so, you could certainly put something together with a microcontroller and discrete logic that would fit in a standard system.
If parallel ports and ISA slots disappear down the road... there will be legacy support for the 10 MBit version of USB for quite a while, and the controller for that is simple enough that you could easily build one with a microcontroller and some glue logic.
In summary, I don't think there will be a problem any time soon.
Are we even to the point when a normal PC could handle Gigabyte? And if so, why not use optical?
A 32-bit 33 MHz PCI bus can support one (1) gigabit ethernet card at full capacity (card's bandwidth is about 100 Mbytes/sec, PCI 32/33 is 133 Mbytes/sec).
If you want to stick multiple cards in (e.g. for a small hypercube-style cluster), buy motherboards that support 64/33 or 64/66 (I was drooling over the dual-processor 64-bit-PCI AMD boards a little while back).
Gigabit ethernet over copper has the advantage of running over your existing cabling (i.e. cat-5 is fine). This avoids having to muck about with fiber, as fiber is a PITA to maintain yourself (getting optically perfect connections for the fiber jacks is picky).
The way gigabit ethernet is encoded on cat-5 cable is both sneaky and elegant.
I want to congratulate the company on making a motherboard that is virtually useless to anybody who isn't bleeding edge. I don't even have a single USB device, and I still use ISA cards extensively because they'er so damn cheap.
I used to do that. Then I decided that I'd rather not have to beat my head against a wall mucking with IRQ conflicts and port addresses to save $10.
USB keyboards are dirt cheap. USB mice are dirt cheap. If you're shelling out for a new system in the first place, replacing keyboards and mice are a negligeable cost (and you'd want new ones regardless, so that you can still keep the old machine active).
Graphics-wise, I'd have to be paid a lot of money to go back to using a graphics card obsolete enough to be ISA, even if all I'm doing is running a 2D desktop. Network-wise, PCI network cards are *almost* as dirt-cheap as your keyboard and mouse.
In summary: If you're buying a new motherboard at all, you can afford to upgrade the peripherals.
The IRS can't require you to be successful as a business.. If you go bankrupt, you get to write it off.. In fact.. Hell, look at Enron.. It even works on a macro scale.
All you 'have to do' is pay your taxes. You can pursue profitability.. But no one will ever require you to be successful.
The way business expenses work up here at least (Canada) is that they don't count towards your taxable income.
You don't get free money - you just don't have to pay as much income tax on the money you _do_ earn.
If you're making little enough not to be paying much tax - which I'll bet you are, if hawking herbal supplements is your primary source of income - then you aren't going to get much back from having less taxable income.
Satellites are one of many redundant networks.
on
Space Wars
·
· Score: 3, Informative
imagine if a foreign country went after our other satellites? Boom, no more cell phone, no more tv, no more satellite internet. You could seriously harm a nation's communications by targeting their satellites. Also, it is not just consumer end stuff, but much of the backbone of the communications go through satellites.
...And through ground-based fiber, and through microwave relays (all those metal towers in the middle of nowhere that you drive past).
Satellites are very useful for sending _small_ amounts of information over long distances to destinations that are relatively isolated. High-bandwidth communications to/from densely populated and well-connected areas don't go through satellites.
Knocking out microwave relay communications would require either a host of *insanely* powerful jammers orbiting overhead, or far more sticks of dynamite than is likely to be practical.
Knocking out fiber communications would involve taking out all of the fiber routing nodes on the continent, or cutting an insanely large number of backbone cables.
Taking out satellites isn't a cakewalk either (it only takes a box of nails, but the box has to be very high up and positioned to within a few metres).
In summary, I think the lower levels of the US's communications network are robust enough to survive virtually all practical attacks (if an enemy can wipe out the communications infrastructure, we have bigger problems than just losing communications).
720x240 at 60 Hz, interlaced to give you 720x480 at 30 Hz, if I remember correctly. Some of the rows/columns aren't visible, though.
Just as a data point, since the exact values aren't terribly relevant:).
Not to mention it's analog.. you may wonder what the difference is, but the fact is that going from analog to digital requires at least 10x more bandwidth. It's simply because analog is much more noise-tolerant... your signal may be affected, but it doesn't result in catastrophic loss as it does in digital systems.
It turns out that this isn't quite correct, for a couple of reasons.
Firstly, there's no reason to transmit the display signal digitally. We've all been using analog CRTs for years without a problem; digital is only required within the computer, where we want to be able to manipulate data without loss. Lossiness on the final output stage is tolerable.
Secondly, it turns out that you can transmit digital signals much more densely than you estimate. A factor of 10 is what I'd expect for one bit per sample plus a little bit of error correction. You can actually get much, much more than this (a 56k modem gets around 4-6 bits per sample, if memory serves). More aggressive error correction codes let you correct for a surprising amount of noise, too.
In short, I think you could do it with only about a factor of 2 bandwidth increase, especially over short range under controllable conditions.
Lastly, you have a vast amount of bandwidth available. If there's enough airspace to transmit 60+ channels of television at relatively low frequencies, finding a window for monitors shouldn't be an unsolvable problem.
Except synchronization based on platform types will fail completely. x86's handle math differently based on their internal arch (athlons, PIV, PIII, 486). They even handle it differently within the same generation.
All you're pointing out is that you can't elminate parameter noise even for an x86-only game. This is why games still do _some_ synchronization:).
Check the X-wing documentary from a couple of weeks ago for a description of how they chose to handle the problem.
Add to that all the processors are processing the data a different speeds.
This actually isn't relevant. All that's required is that if, say, you're sending an update every twentieth of a second in virtual time, all of your machines be able to update relevant parts of their world-models in less than one twentieth of a second.
In practice, it's even more flexible than this, as the update time steps for each client instance are allowed to be arbitrary. Naturally, this introduces more noise due to roundoff errors not maching up.
There is no way to avoid having to exploit this effect. A full update of the visible game state for each user would take a vast amount of bandwidth to send every frame. All you've shown here is that _some_ drift is unavoidable under real conditions - and I agree. The problem is that porting across architectures can only make the problem worse (and thus increase the cost of managing the problem).
Umm why do you think that, if I do a scientific calculation on my Intel based system and run the same computations on my Sparc I get the same results, it's called IEEE floats.
An IEEE-compliant math unit will produce an answer that is correct to the ideal value within half a unit-last-place.
Except for operations like square roots, logarithms, sine/cosine, and so forth, where it's impractical or impossible to implement this degree of precision for all possible cases.
And except for systems where the math units aren't perfectly IEEE-compliant (there are many).
So while in theory you could do this, in practice you'll get small errors.
And this isn't even touching the fact that things like dot products are very sensitive to the order in which you choose to perform the fundamental operations required to implement them. If the compilers for your different platforms choose slightly different implementations, you have another source of noise.
In summary, the problem definitely exists.
Besides most of the game data is int's not float. For instance why would you need a fraction of a hit point or a fraction of a movement point, pen and paper games use just int's for instance.
Any game that has a true-3D world uses float to represent pretty much all coordinates. It doesn't take any more space to store a single-precision float than a 32-bit int, and it saves a lot of conversion hassle.
Surely it all depends on what the client side does? If all you end up doing is sending character data, co-ordinates etc why should it be particularily difficult?
The way real-time game synchronization usually works is to take advantage of the fact that copies of the game on identical platforms will behave exactly the same way, given identical input. This lets you only transmit a very small number of state changes between machines in a multi-player game, with the automatic update of the rest of the world keeping most things in sync.
This starts to break down when you have platforms that handle math substantially differently (x86 and MIPS, in this case). Your world-update calculations will produce slightly different results, which will evolve into very large changes if you don't send corrective information every so often.
You can build a game to be resistant to this kind of drift, and to correct this kind of drift, but the usual result is that you have to send considerably more data during updates, with the game possibly becoming more sensitive to latency as well.
That would quickly become a nightmare the moment they decided to release a patch for the PC version. Unless they are going to require you to add a hard drive to your PS2 as well, there is no way they are going to be able to patch a PS2 version to match.
Actually, you could get around this by having the game contact Sony for an in-memory patch on startup. Keep updating the patch on the master server, and the game patching takes care of itself.
You'd just have to patch a bit more intelligently than most games do (game patches don't *have* to be megabytes if you're only fixing bugs in the engine code).
The test ban was enacted so that nations would STOP designing better planet-busters. Now we have shown that it is possible for people to design nukes in thier basement (assuming their basement has a 12 teraflop computer).
Should we feel any more secure knowing that India and Pakistan can now quietly design better atomic arsenals to annihilate each other with?
Unfortunately, there is no way to stop people from being able to perform computing simulations like this without also severely limiting most of their other technology - which would be grossly unethical. This is the same kind of problem as the old "limit the knowledge of how to build nuclear weapons" thread that was going around a few years back; to make it impossible for anyone to figure out how to make nuclear weapons, you have to basically condemn them to an early-20th-century knowledge of science forever.
It's more practical (and more convenient from an ethical standpoint too) just to look for signs of nuclear weapon production. It takes quite a bit of industry to refine the required materials; this can be detected if the watchers are vigilant. You can also detect the required nuclear plants from orbit with the right kind of sensors and a bit of patience (and it would surprise me greatly if the US didn't already have a host of satellites quietly looking for gamma ray glow on the ground).
Limiting the ability to *design* nuclear weapons also doesn't really limit a nation's ability to *get* nuclear weapons, so I'd argue that the purpose of the test ban treaty is more to prevent escalation between the existing nuclear powers than to prevent new people from gaining nuclear capability.
I find various new games fun and many old games fun, but there is something different about the new games.. theres just a vague feeling that something is "missing", or rather, that the old games just had something that the new games don't. Its subtle and I can't quite put my finger on what it is that the old games had
This sounds like nostalgia to me:).
I felt this way once, and dusted off my old copy of Populous. And was appalled at how horrid and difficult to use the interface was, among other things. The game was still ok, but hardly the playground of infinite fun that I'd remembered.
Ditto when I dusted off the TRS-80 I programmed on as a kid.
Ditto when I tried playing Warcraft I again.
I don't believe that new games are fundamentally worse than old games. I get that "spark of wonder" feeling thinking about some of my older game escapades... but then I think back a year later and get the same feeling about the games that I was considering mundane when I was first waxing nostalgic. This "good old days" feeling seems to be something that my mind adds after the fact.
Thus, I am doubly doubtful of the "old games just *felt* better" argument, as I've been seduced by it myself.
This doesn't mean old games aren't fun; I still play Civ. It just means they weren't any more magical than current games.
Wouldn't it make more sense to use a proton or a positively-charged ion, so that you could easily hold it in one place?
Anyone here familiar with the actual experiment being proposed, who can clue me in? (might just be looking for a double event from a radioactive decay, or some similar measuring trick)
there is currently a trend in the computer game industry where graphical realism is considered a suitable substitute for creativity. Most people posting on this board seem to be too young to even be aware of what games actually used to be like, so they don't really understand it, but basically with the games of the 80's and early 90's the developers were essentially "forced" to be very creative in devising interesting, abstract gameplay strategies, simply because realism wasn't an option.
I think you underestimate the memories and/or age of most of the people here.
I further think that you're just *remembering* only the good games. Garbage games have been around for as long as there have been platforms to write them for. Read up on Seanbaby's "20 worst atari games" feature for an example.
Crap games certainly exist. See my original post. Before garbage could be marketed based on polycount, it was marketed based on which movie they'd bought the rights to. Programmers and (especially) gaming companies have never been "forced" to write good games, no matter how restrictive a platform they've been developing for.
In summary, I don't buy the "old games were better" argument.
I'm still trying to figure out what people are up in arms about.
Realism helps a lot for some types of game, and doesn't help for others. A wise team will use realism when and only when it's useful.
Where's the problem?
Marketing will try to pitch crap based on its technical merits, but this has been true for longer than most of us have been born. "realism" is just the latest buzzword. How is this "realism"'s fault?
This would be a Bad Idea in embedded devices, because they may very well be designed not to be upgraded.
This would also be a Bad Idea in any installation where the person maintaining a machine may change (which covers just about everywhere). It's hard enough keeping track of everything on your own machine - what about a machine you inherit from a previous administrator?
The machine suddenly stops doing something vital when the software expires, and you have to track down what and where it was.
Better just to write "review the installed version of Widget X" in your day planner at regular intervals.
Why should this guy have to go through any trouble at all? He wasn't wearing anything that could be construed as a bomb or a weapon of any kind.
Unless he's changed it very recently, his gear looks like a fanny pack filled with gutted computer parts, with misc. cables going out to various peripherals, many with visible PCBs and so forth.
He may have cleaned it up a bit, but take this and add a reasonable-sized battery, and you have a rig that looks a lot like your "ACME Personal Bomb" from any action movie from the past decade or two.
Add to this the fact that Prof. Mann is a bit on the eccentric side and that he would very likely have gotten pushy with the guards when they challenged him (trust me on this one), and what you have is a recipe for a really bad day (and a really golden publicity opportunity, which was probably the plan).
Slashdot Mirror
SGI markets 1024 machines "out of the box" and there is a 6000+ CPU machine on the top 10. They are currently working on the 2nd generation which will allow 33,000 CPUs (15 bits of CPU's). It'll be cool.
My mistake re. SGI; thanks for making me aware of this (another poster pointed it out too).
Also, WRT address space, I would think memory access on these things is quite heavily abstracted for any userland tasks. When you reach outside of any one machine on the cluster, conventional memory access methods probably go out the window anyway. The ASCI Red was just a bunch of P6's soldered together after all, and it doesn't seem to be having too many problems.
I seriously doubt that any x86 cluster uses a unified address space from any given task's point of view.
Abstracting memory accesses could let you make your real address space a window into a larger one, but that would have some pretty nasty overhead.
IBM RS/6000's with Power4 cores?
:). Knew I'd forgotten something.
Whoops
Re: your less-than-insightful comment on x86: Intel's ASCI Red has 9472 x86 CPUs. Guess what - they don't share 4GB memory...
*sigh*.
If you're dealing with problems where you don't need to have the entire data set visible to all processors, great; use x86.
If you need to map the entire address space, you need more than the 36 or so bits that x86 offers you.
NUMA. Go look it up. :-)
NUMA doesn't touch the address space problem, or the processor-type problem. It's just a way of arranging the memory hierarchy.
They're awfully confident of McKinley not following in the footsteps of Merced if they've placed this order.
This raises an interesting question, though. If you want to build a high-performance compute cluster nowadays... what do you build it out of? The old answer, Alpha, doesn't really apply any more.
Sun is optimized for communications bandwidth, not FLOPS, and I'm not sure if SGI even _offers_ machines that huge. HP is betting on IA64. And x86 is competely unsuitable, for memory space reasons if nothing else.
What am I missing?
It's easier to make a funny thing with a cheap Motorola 6800 or a Zilog Z80 than with a Intel586 or AMD K7. Both for the hardware side (it's only 40 pins and 2MHz) as for the software side (just a couple of registers).
You can buy a host of programmable microcontrollers from a variety of vendors; check Digikey's catalog for a sampling. Many of these should adequately substitute for a Z80.
Also, how "easy" is it these days to add an self-developped extensionboard into your computer?
Not that hard.
I was building a project driven off of a parallel port a couple of weeks back. These won't go away for a few years yet, and you can clock them as slowly as you want to.
You can also still find motherboards with ISA slots for new machines; at 8 MHz or so, you could certainly put something together with a microcontroller and discrete logic that would fit in a standard system.
If parallel ports and ISA slots disappear down the road... there will be legacy support for the 10 MBit version of USB for quite a while, and the controller for that is simple enough that you could easily build one with a microcontroller and some glue logic.
In summary, I don't think there will be a problem any time soon.
Are we even to the point when a normal PC could handle Gigabyte? And if so, why not use optical?
A 32-bit 33 MHz PCI bus can support one (1) gigabit ethernet card at full capacity (card's bandwidth is about 100 Mbytes/sec, PCI 32/33 is 133 Mbytes/sec).
If you want to stick multiple cards in (e.g. for a small hypercube-style cluster), buy motherboards that support 64/33 or 64/66 (I was drooling over the dual-processor 64-bit-PCI AMD boards a little while back).
Gigabit ethernet over copper has the advantage of running over your existing cabling (i.e. cat-5 is fine). This avoids having to muck about with fiber, as fiber is a PITA to maintain yourself (getting optically perfect connections for the fiber jacks is picky).
The way gigabit ethernet is encoded on cat-5 cable is both sneaky and elegant.
apparently Pricewatch.com has D-Link 8-port 10/100/1000baseT auto-detect switches listed for under $150!
;).
These are for 8x100-base-T with a gigabit uplink. I researched this a while ago, when speccing out my dream network
The cheapest full-gigabit switch D-link sells is about $1500.
I want to congratulate the company on making a motherboard that is virtually useless to anybody who isn't bleeding edge. I don't even have a single USB device, and I still use ISA cards extensively because they'er so damn cheap.
I used to do that. Then I decided that I'd rather not have to beat my head against a wall mucking with IRQ conflicts and port addresses to save $10.
USB keyboards are dirt cheap. USB mice are dirt cheap. If you're shelling out for a new system in the first place, replacing keyboards and mice are a negligeable cost (and you'd want new ones regardless, so that you can still keep the old machine active).
Graphics-wise, I'd have to be paid a lot of money to go back to using a graphics card obsolete enough to be ISA, even if all I'm doing is running a 2D desktop. Network-wise, PCI network cards are *almost* as dirt-cheap as your keyboard and mouse.
In summary: If you're buying a new motherboard at all, you can afford to upgrade the peripherals.
The IRS can't require you to be successful as a business.. If you go bankrupt, you get to write it off.. In fact.. Hell, look at Enron.. It even works on a macro scale.
All you 'have to do' is pay your taxes. You can pursue profitability.. But no one will ever require you to be successful.
The way business expenses work up here at least (Canada) is that they don't count towards your taxable income.
You don't get free money - you just don't have to pay as much income tax on the money you _do_ earn.
If you're making little enough not to be paying much tax - which I'll bet you are, if hawking herbal supplements is your primary source of income - then you aren't going to get much back from having less taxable income.
imagine if a foreign country went after our other satellites? Boom, no more cell phone, no more tv, no more satellite internet. You could seriously harm a nation's communications by targeting their satellites. Also, it is not just consumer end stuff, but much of the backbone of the communications go through satellites.
...And through ground-based fiber, and through microwave relays (all those metal towers in the middle of nowhere that you drive past).
Satellites are very useful for sending _small_ amounts of information over long distances to destinations that are relatively isolated. High-bandwidth communications to/from densely populated and well-connected areas don't go through satellites.
Knocking out microwave relay communications would require either a host of *insanely* powerful jammers orbiting overhead, or far more sticks of dynamite than is likely to be practical.
Knocking out fiber communications would involve taking out all of the fiber routing nodes on the continent, or cutting an insanely large number of backbone cables.
Taking out satellites isn't a cakewalk either (it only takes a box of nails, but the box has to be very high up and positioned to within a few metres).
In summary, I think the lower levels of the US's communications network are robust enough to survive virtually all practical attacks (if an enemy can wipe out the communications infrastructure, we have bigger problems than just losing communications).
Actually, TV is more along 300 lines I think.
:).
720x240 at 60 Hz, interlaced to give you 720x480 at 30 Hz, if I remember correctly. Some of the rows/columns aren't visible, though.
Just as a data point, since the exact values aren't terribly relevant
Not to mention it's analog.. you may wonder what the difference is, but the fact is that going from analog to digital requires at least 10x more bandwidth. It's simply because analog is much more noise-tolerant... your signal may be affected, but it doesn't result in catastrophic loss as it does in digital systems.
It turns out that this isn't quite correct, for a couple of reasons.
Firstly, there's no reason to transmit the display signal digitally. We've all been using analog CRTs for years without a problem; digital is only required within the computer, where we want to be able to manipulate data without loss. Lossiness on the final output stage is tolerable.
Secondly, it turns out that you can transmit digital signals much more densely than you estimate. A factor of 10 is what I'd expect for one bit per sample plus a little bit of error correction. You can actually get much, much more than this (a 56k modem gets around 4-6 bits per sample, if memory serves). More aggressive error correction codes let you correct for a surprising amount of noise, too.
In short, I think you could do it with only about a factor of 2 bandwidth increase, especially over short range under controllable conditions.
Lastly, you have a vast amount of bandwidth available. If there's enough airspace to transmit 60+ channels of television at relatively low frequencies, finding a window for monitors shouldn't be an unsolvable problem.
Except synchronization based on platform types will fail completely. x86's handle math differently based on their internal arch (athlons, PIV, PIII, 486). They even handle it differently within the same generation.
:).
All you're pointing out is that you can't elminate parameter noise even for an x86-only game. This is why games still do _some_ synchronization
Check the X-wing documentary from a couple of weeks ago for a description of how they chose to handle the problem.
Add to that all the processors are processing the data a different speeds.
This actually isn't relevant. All that's required is that if, say, you're sending an update every twentieth of a second in virtual time, all of your machines be able to update relevant parts of their world-models in less than one twentieth of a second.
In practice, it's even more flexible than this, as the update time steps for each client instance are allowed to be arbitrary. Naturally, this introduces more noise due to roundoff errors not maching up.
There is no way to avoid having to exploit this effect. A full update of the visible game state for each user would take a vast amount of bandwidth to send every frame. All you've shown here is that _some_ drift is unavoidable under real conditions - and I agree. The problem is that porting across architectures can only make the problem worse (and thus increase the cost of managing the problem).
Umm why do you think that, if I do a scientific calculation on my Intel based system and run the same computations on my Sparc I get the same results, it's called IEEE floats.
An IEEE-compliant math unit will produce an answer that is correct to the ideal value within half a unit-last-place.
Except for operations like square roots, logarithms, sine/cosine, and so forth, where it's impractical or impossible to implement this degree of precision for all possible cases.
And except for systems where the math units aren't perfectly IEEE-compliant (there are many).
So while in theory you could do this, in practice you'll get small errors.
And this isn't even touching the fact that things like dot products are very sensitive to the order in which you choose to perform the fundamental operations required to implement them. If the compilers for your different platforms choose slightly different implementations, you have another source of noise.
In summary, the problem definitely exists.
Besides most of the game data is int's not float. For instance why would you need a fraction of a hit point or a fraction of a movement point, pen and paper games use just int's for instance.
Any game that has a true-3D world uses float to represent pretty much all coordinates. It doesn't take any more space to store a single-precision float than a 32-bit int, and it saves a lot of conversion hassle.
Surely it all depends on what the client side does? If all you end up doing is sending character data, co-ordinates etc why should it be particularily difficult?
The way real-time game synchronization usually works is to take advantage of the fact that copies of the game on identical platforms will behave exactly the same way, given identical input. This lets you only transmit a very small number of state changes between machines in a multi-player game, with the automatic update of the rest of the world keeping most things in sync.
This starts to break down when you have platforms that handle math substantially differently (x86 and MIPS, in this case). Your world-update calculations will produce slightly different results, which will evolve into very large changes if you don't send corrective information every so often.
You can build a game to be resistant to this kind of drift, and to correct this kind of drift, but the usual result is that you have to send considerably more data during updates, with the game possibly becoming more sensitive to latency as well.
That would quickly become a nightmare the moment they decided to release a patch for the PC version. Unless they are going to require you to add a hard drive to your PS2 as well, there is no way they are going to be able to patch a PS2 version to match.
Actually, you could get around this by having the game contact Sony for an in-memory patch on startup. Keep updating the patch on the master server, and the game patching takes care of itself.
You'd just have to patch a bit more intelligently than most games do (game patches don't *have* to be megabytes if you're only fixing bugs in the engine code).
The test ban was enacted so that nations would STOP designing better planet-busters. Now we have shown that it is possible for people to design nukes in thier basement (assuming their basement has a 12 teraflop computer).
Should we feel any more secure knowing that India and Pakistan can now quietly design better atomic arsenals to annihilate each other with?
Unfortunately, there is no way to stop people from being able to perform computing simulations like this without also severely limiting most of their other technology - which would be grossly unethical. This is the same kind of problem as the old "limit the knowledge of how to build nuclear weapons" thread that was going around a few years back; to make it impossible for anyone to figure out how to make nuclear weapons, you have to basically condemn them to an early-20th-century knowledge of science forever.
It's more practical (and more convenient from an ethical standpoint too) just to look for signs of nuclear weapon production. It takes quite a bit of industry to refine the required materials; this can be detected if the watchers are vigilant. You can also detect the required nuclear plants from orbit with the right kind of sensors and a bit of patience (and it would surprise me greatly if the US didn't already have a host of satellites quietly looking for gamma ray glow on the ground).
Limiting the ability to *design* nuclear weapons also doesn't really limit a nation's ability to *get* nuclear weapons, so I'd argue that the purpose of the test ban treaty is more to prevent escalation between the existing nuclear powers than to prevent new people from gaining nuclear capability.
I find various new games fun and many old games fun, but there is something different about the new games .. theres just a vague feeling that something is "missing", or rather, that the old games just had something that the new games don't. Its subtle and I can't quite put my finger on what it is that the old games had
:).
This sounds like nostalgia to me
I felt this way once, and dusted off my old copy of Populous. And was appalled at how horrid and difficult to use the interface was, among other things. The game was still ok, but hardly the playground of infinite fun that I'd remembered.
Ditto when I dusted off the TRS-80 I programmed on as a kid.
Ditto when I tried playing Warcraft I again.
I don't believe that new games are fundamentally worse than old games. I get that "spark of wonder" feeling thinking about some of my older game escapades... but then I think back a year later and get the same feeling about the games that I was considering mundane when I was first waxing nostalgic. This "good old days" feeling seems to be something that my mind adds after the fact.
Thus, I am doubly doubtful of the "old games just *felt* better" argument, as I've been seduced by it myself.
This doesn't mean old games aren't fun; I still play Civ. It just means they weren't any more magical than current games.
Why a neutron?
Wouldn't it make more sense to use a proton or a positively-charged ion, so that you could easily hold it in one place?
Anyone here familiar with the actual experiment being proposed, who can clue me in? (might just be looking for a double event from a radioactive decay, or some similar measuring trick)
there is currently a trend in the computer game industry where graphical realism is considered a suitable substitute for creativity. Most people posting on this board seem to be too young to even be aware of what games actually used to be like, so they don't really understand it, but basically with the games of the 80's and early 90's the developers were essentially "forced" to be very creative in devising interesting, abstract gameplay strategies, simply because realism wasn't an option.
I think you underestimate the memories and/or age of most of the people here.
I further think that you're just *remembering* only the good games. Garbage games have been around for as long as there have been platforms to write them for. Read up on Seanbaby's "20 worst atari games" feature for an example.
Crap games certainly exist. See my original post. Before garbage could be marketed based on polycount, it was marketed based on which movie they'd bought the rights to. Programmers and (especially) gaming companies have never been "forced" to write good games, no matter how restrictive a platform they've been developing for.
In summary, I don't buy the "old games were better" argument.
I'm still trying to figure out what people are up in arms about.
Realism helps a lot for some types of game, and doesn't help for others. A wise team will use realism when and only when it's useful.
Where's the problem?
Marketing will try to pitch crap based on its technical merits, but this has been true for longer than most of us have been born. "realism" is just the latest buzzword. How is this "realism"'s fault?
Play what you find fun. The market will follow.
Um, you might want to actually read about the satellite before assuming it uses radiothermal generators.
The great big solar panels in the picture of the satellite might have been a hint that it didn't use nuclear power.
From the HETE pages (describing HETE-2, an exact duplicate of the HETE-1 craft whose launch was unsuccessful):
The HETE-2 power system hardware consists of
You can find more information on the specs of the HETE satellites at http://space.mit.edu/HETE/spacecraft.html .
This would be a Bad Idea in embedded devices, because they may very well be designed not to be upgraded.
This would also be a Bad Idea in any installation where the person maintaining a machine may change (which covers just about everywhere). It's hard enough keeping track of everything on your own machine - what about a machine you inherit from a previous administrator?
The machine suddenly stops doing something vital when the software expires, and you have to track down what and where it was.
Better just to write "review the installed version of Widget X" in your day planner at regular intervals.
Why should this guy have to go through any trouble at all? He wasn't wearing anything that could be construed as a bomb or a weapon of any kind.
Unless he's changed it very recently, his gear looks like a fanny pack filled with gutted computer parts, with misc. cables going out to various peripherals, many with visible PCBs and so forth.
He may have cleaned it up a bit, but take this and add a reasonable-sized battery, and you have a rig that looks a lot like your "ACME Personal Bomb" from any action movie from the past decade or two.
Add to this the fact that Prof. Mann is a bit on the eccentric side and that he would very likely have gotten pushy with the guards when they challenged him (trust me on this one), and what you have is a recipe for a really bad day (and a really golden publicity opportunity, which was probably the plan).