All you need is a faraday cage, no? I would have thought that gvmt buildings mandated those built into the walls.
For a personal computer, just wrapping the whole thing in fine chicken wire should do it, I would have thought.
Of course, as the recent drug trial shows, unless you have maintain physical access security of your systems at all times, a dedicated attacker can just install a keyboard snooper on the cable itself, making tempest unnecessary.
Assuming that the film advances at a constant rate, the second or third shutters would show half of one frame and half of the next (unless you did some fancy re-orientation optics so you reposition the frames on the screen, which I highly doubt).
I guess you could just have several lenses, but that doesn't seem right either.
so, please expound on how you get multiple refreshes of one frame in a mechinical projector.
I think a swedish Viggen is the only plane ever to get radar lock on one (after which you can fire your missle and head home). Bit of a fluke, but still. The thing about SR-71s is that they are pretty much build exclusively for invading other countries airspace and to be fast enough to get away with it.
I'd imagine that the advent of hi-res satelite photography -- if ikonos is getting better than 1m, what don't you think the spooks can do? -- coupled with high operating costs and bad PR were instrumental in its decommision.
does all of it need to be on WORM, or can you just use that to store several hashes? I'm suggesting you use normal TAPE to store the backups and then store a hash on WORM. The hash authenticates the non-tamper resistant tape, while remaining cheap and convenient since you only need one.
Howabout a hybrid solution: I store the files as normal backups, on tape/disk/whatever, but store a secture hash on the worm drive. So I use the worm capabilities to provide the authentication, but the price-per-bit of magnetic drives for the data.
hrm?
On a side note, I seem to recall plan-9 using a worm to do its nightly backups. However, since they were all incremental, they didn't need much space -- I recall Richie (I think) musing that storage capacity was growing faster than they were using it...
The question was aimed at delving into microsoft's development practices that seem to inevitably produce insecure-by-default products.
The answer was about how quick they are at fixing/patching/disabling the holes as they are publicised.
Yes the patch came out and fixed that hole, but that was just one example. The question was why there seems to be so many, as if they are put in on purpose.
When the answer doesn't address the question, that is a dodge. Plain and simple.
I wonder; would it be more useful to market these as reprogrammable CPUs? Ie, don't make the poor hardware designer design the whole CPU, but give them a few instructions that you'll take care of the decoding and commit-in-order and speculation, but they get to design the actual instruction.
Outlinish they'd declare: this instruction reads registers x,y,z, writes a,b,c, and will require so many cycles to complete after inputs.
Has this been tried and failed, is this what they do, or are there other reasons why It Would Never Work?
One area where these babies ought to shine is in modern ciphers systems. DES, twofish, as DSPs, maybe even as CODECs (tho that might be a bit much for the gatebudget)... anything where you can partially evaluate the program into hardware. This would work very well on a daughter card, where you can set up DMA controllers to just pump data in and out of the FPGA.
The hard part is designing the circuit. Compilation down to silicon is a known hard job, with layout and drawing abstraction boundaries being two main stumbling blocks.
blue sky musing: On the horizon for mainstream acceptance are profiling feedback optimisers, which produce specialised versions of code that run very fast for a limited set of [common] inputs. These currently go from a higher level language to a lower level language (java JITs like HotSpot or transmeta's codemorphing) or from the lowlevel language to the same lowlevel language (HPs dynamo).
It would be really cool to see this technology applied to creating FPGAs, where the meta software notices that a certain basic block is taken often, and has mainly bittwidling operations. If it is taken often enough, and is long enough (this is where the specialisation of dynamo comes in -- it basically just creates optimised long basic blocks) it makes perfect sense to compile it to silicon.
Eventually, the Ghz race WILL pewter out, and we'll be forced to this sort of generalised specialisation for getting the 90/10 any faster.
ugh. AND if you have a case of A covers B covers C covers A (which you can set up with three triangles) you can get a situation that requires splitting (what I called decimation) but has no intersection.
so painter's (didn't know it was called that) isn't even able to deal with the simple cases.
You'll need a vacuum too. To get 50hz, you'll need a 300rpm rotation. If it's a helix, you'll be stuck with a gigantic propellor, and if it is flat, you have a centripetal pump. So you need vacuum.
(side note: couldn't we just use one of those vapor-trail detectors and two or three low power lasers to acheive 3-d vector display? The idea being that neither beam would be powerful enough on its own to ionize the vapor, but at their intersection, they would combine to have enough power -- or do we run into the limitation that you can't add quanta?)
Also, both display technolgies have the drawback of being non-occluding. You can't display solid objects, because the front face will be transparent, letting the back shine through.
This will of course be fine for air traffic control, and for things like displaying MRI scans doctors are already proficient at reading layered data, but it will be a hassle for many potential uses.
Of course, the ability to select what is displayed will help.
It is not clear exactly how much control you have of the system and where you want this software to take over.
However, the fig2dev program is able to translate fig files (generated by xfig) to any number of formats -- eps among them. The fig file format is pretty straight forward vector graphics, so all you need to do is decimate all the surfaces to triangles, rotate the figure, sort the surfaces in z-depth order, and dump it to fig format.
fig2dev makes an eps file, and then you let the printer take care of it. Filled polygons is something a postscript printer can draw in its sleep; hidden surface removal becomes moot.
Someone already suggested the OpenGl -> PS library, but my suggestion has the benefit of requiring almost no APIs at all; just rotate, decimate, sort, and dump to file.
bit of my favorite hooby horse, so I hope you'll excuse me repeating it:
ALL key-based encryption is in NP (it is easy to verify a suggested key). So it matters little whether it is based on factorisation, points on elliptic curve, rotor positions, or S boxes.
BTW I do believe that while nobody knows whether FACTOR is NP, it is commonly held NOT to be NP COMPLETE.
Besides, the copper that some chips are made of is a pretty good conductor of heat already. Perhaps all that we need are copper rods embedded in chips (cunningly placed to avoid short circuits, and equally cunningly shaped to avoid inductive effects on the useful copper) to help draw out the heat?
But then I think that if I had actually read the article, I might have something more insightfull to say.
Since the originals are of much lower quality than movies, that lowers the bandwidth requirements even more. Are we looking at 150MB for an hour's worth of entertainment?
I personally don't have a television, but I do have sneakernet fast internet access with my laptop, so if I were able to rent/download/subscribe to just-past-their-prime television series to play in the BG while I cook, I'd be all over that.
However, the price has to be right. $10 a month gets me basic cable, and $200 gets me an All-In-Wonder to skip comercials with, so they have to compete with those prices.
I wonder whether bandwidth prices for the studios can make that worth their while.
Thing is, I read your comment before I read the article, and I'm thinking to myself, "Cool, berlin has their display-postscript driver" --which I assumed existed because someone else mentioned NeWS -- "implemented in forth?! Nifty".
I was more than a little bummed to find out I was somewhat off.
It would take a grand total of four apps ported to berlin to make it a viable alternative for me.
1) emacs
2) xdvi/xdpf (they wouldn't be called that of course)
3) xterm
4) netscape (or the frankly superior Internet Explorer. I wish it weren't so, but t'is.)
Those are the only gui apps that I've used during the daylight hours for the last four odd years. Incidentally, they are also the apps that would benefit most from anti-aliasing (preferably sub-pixel).
The article didn't quite make it clear to me whether an X compatiblity layer would be possible, either at the developer level (recompile all apps) or the user level (just start the Xf86 server that hooks up with a running berlin server). Anyone want to chip in with an opinion?
Also, (somewhat OT): does anyone have a status update for Xrender in the newest Xfree86? I seem to recall a page promising sub-pixel anti-aliasing for my laptop, and I want it badly. That would make my emacs SOO easy to read.
The recent release was mum tho, so I'm sorta in the dark.
I'm suprised that this sort of thing isn't better done in orbit, where at least gravity and temparature are on your side. You could imagine slipping a few big mirrors up to the lagrange points, and focussing them on hubble or something for the inferomtery.
That must be the first time the lameness filter actually filtered junk!
:-P
seriously, that lameness filter has pissed me off so many times (ever spend 15 minutes drawing ascii art to illustrate a point and then...) that I'm kinda glad someone else has run into it as well. If they wanna keep it for ACs I'd understand that, but people with karma should be allowed to bypass it, IMHO.
Re:Multi Processors under Win9x
on
Emergence of SMT
·
· Score: 1
well, the best(*) a 100% parallel system could do in that configuration would be to equal the faster chip. 2 x 500 = 1 x 1000 !
Now if the article had said that a 2 x 800 was just barely as fast as a 1 x 1000, under an OS with the capability to use both CPUs, that would have been noteworthy. But this must have been one of the least informative comparisons I've heard in a long time.
(*) disregarding memory bandwidth, which depends on busses and what not.
I was under the impression that we were coming out of a small ice age. This is claimed as evidence for winters being whiter in your grandparent's days.
This is often claimed as the real reason why the earth is warming up, as opposed to greenhouse effects, which are presumed to take longer to gestate. (I must admit that I had thought it strange that we had already had an effect on the climate -- which is not to say that our current emissions won't wreack havoc in the next centuries).
I'm reading the book with Feyman's talks and lectures (some of them) called "the joy of finding stuff out" or something similar, and he also mention the work of eacly 20 century mathematician... forget the name... Hilbert? Anyways, this guy went off and worked out all these therories about how vectors and matricies that appeared to have no use whatsoever. Until quantuum mechanics came around, and Pauli had to reinvent much of it.
Now Pauli was a smart guy, so a lesser intellect might have looked up Hilbert's stuff rather than reinventing it.
This is an instance of the old ATM vs IP or CISC vs RISC debates. It's the old engineering tradeoff: work smart but slow or work quick and dirty. Tile based rending is an instance of smart and slow, ie they do no more work than they have to, and thus get away with slower clocks and memory. The NVIDIA card is quick and dirty.
Historically, it is almost always the case that quick and dirty is the cheaper way to go, as it allows economies of scale to come into play. However, it is seeming more and more like the memory bandwidth bottleneck is here to stay, so the smart and slow approach is looking pretty good. Likewise as we run into physical limitations for network bandwidth, IP is going to have a harder and harder time to provide acceptable QoS and multicast solutions and ATM-like technologies will start becoming more prevalent.
napster I thought supported whatever format people saved their songs as. I was under the impression (never tried it, usenet for me) that napster didn't actual create (or until recently) screen any files.
Slashdot Mirror
$5K?
All you need is a faraday cage, no? I would have thought that gvmt buildings mandated those built into the walls.
For a personal computer, just wrapping the whole thing in fine chicken wire should do it, I would have thought.
Of course, as the recent drug trial shows, unless you have maintain physical access security of your systems at all times, a dedicated attacker can just install a keyboard snooper on the cable itself, making tempest unnecessary.
How does THAT work?
Assuming that the film advances at a constant rate, the second or third shutters would show half of one frame and half of the next (unless you did some fancy re-orientation optics so you reposition the frames on the screen, which I highly doubt).
I guess you could just have several lenses, but that doesn't seem right either.
so, please expound on how you get multiple refreshes of one frame in a mechinical projector.
I think a swedish Viggen is the only plane ever to get radar lock on one (after which you can fire your missle and head home). Bit of a fluke, but still. The thing about SR-71s is that they are pretty much build exclusively for invading other countries airspace and to be fast enough to get away with it.
I'd imagine that the advent of hi-res satelite photography -- if ikonos is getting better than 1m, what don't you think the spooks can do? -- coupled with high operating costs and bad PR were instrumental in its decommision.
I posted this above, but here is a better place:
does all of it need to be on WORM, or can you just use that to store several hashes? I'm suggesting you use normal TAPE to store the backups and then store a hash on WORM. The hash authenticates the non-tamper resistant tape, while remaining cheap and convenient since you only need one.
Speaking of holding up in court,
Howabout a hybrid solution: I store the files as normal backups, on tape/disk/whatever, but store a secture hash on the worm drive. So I use the worm capabilities to provide the authentication, but the price-per-bit of magnetic drives for the data.
hrm?
On a side note, I seem to recall plan-9 using a worm to do its nightly backups. However, since they were all incremental, they didn't need much space -- I recall Richie (I think) musing that storage capacity was growing faster than they were using it...
The question was aimed at delving into microsoft's development practices that seem to inevitably produce insecure-by-default products.
The answer was about how quick they are at fixing/patching/disabling the holes as they are publicised.
Yes the patch came out and fixed that hole, but that was just one example. The question was why there seems to be so many, as if they are put in on purpose.
When the answer doesn't address the question, that is a dodge. Plain and simple.
neato.
I wonder; would it be more useful to market these as reprogrammable CPUs? Ie, don't make the poor hardware designer design the whole CPU, but give them a few instructions that you'll take care of the decoding and commit-in-order and speculation, but they get to design the actual instruction.
Outlinish they'd declare: this instruction reads registers x,y,z, writes a,b,c, and will require so many cycles to complete after inputs.
Has this been tried and failed, is this what they do, or are there other reasons why It Would Never Work?
One area where these babies ought to shine is in modern ciphers systems. DES, twofish, as DSPs, maybe even as CODECs (tho that might be a bit much for the gatebudget)... anything where you can partially evaluate the program into hardware. This would work very well on a daughter card, where you can set up DMA controllers to just pump data in and out of the FPGA.
The hard part is designing the circuit. Compilation down to silicon is a known hard job, with layout and drawing abstraction boundaries being two main stumbling blocks.
blue sky musing: On the horizon for mainstream acceptance are profiling feedback optimisers, which produce specialised versions of code that run very fast for a limited set of [common] inputs. These currently go from a higher level language to a lower level language (java JITs like HotSpot or transmeta's codemorphing) or from the lowlevel language to the same lowlevel language (HPs dynamo).
It would be really cool to see this technology applied to creating FPGAs, where the meta software notices that a certain basic block is taken often, and has mainly bittwidling operations. If it is taken often enough, and is long enough (this is where the specialisation of dynamo comes in -- it basically just creates optimised long basic blocks) it makes perfect sense to compile it to silicon.
Eventually, the Ghz race WILL pewter out, and we'll be forced to this sort of generalised specialisation for getting the 90/10 any faster.
what impresses me the most is the time and effort that went into bypassing the lameness filter.
misguided yes, but you have to give a nod to dedication when you see it. anyone know what software did the ascii-artification?
ugh. AND if you have a case of A covers B covers C covers A (which you can set up with three triangles) you can get a situation that requires splitting (what I called decimation) but has no intersection.
so painter's (didn't know it was called that) isn't even able to deal with the simple cases.
crap.
obl. disc. site is hosed, so I'm speculating:
You'll need a vacuum too. To get 50hz, you'll need a 300rpm rotation. If it's a helix, you'll be stuck with a gigantic propellor, and if it is flat, you have a centripetal pump. So you need vacuum.
(side note: couldn't we just use one of those vapor-trail detectors and two or three low power lasers to acheive 3-d vector display? The idea being that neither beam would be powerful enough on its own to ionize the vapor, but at their intersection, they would combine to have enough power -- or do we run into the limitation that you can't add quanta?)
Also, both display technolgies have the drawback of being non-occluding. You can't display solid objects, because the front face will be transparent, letting the back shine through.
This will of course be fine for air traffic control, and for things like displaying MRI scans doctors are already proficient at reading layered data, but it will be a hassle for many potential uses.
Of course, the ability to select what is displayed will help.
But not precise enough to keep orientation for more than a few minutes? Why is that?
It is not clear exactly how much control you have of the system and where you want this software to take over.
However, the fig2dev program is able to translate fig files (generated by xfig) to any number of formats -- eps among them. The fig file format is pretty straight forward vector graphics, so all you need to do is decimate all the surfaces to triangles, rotate the figure, sort the surfaces in z-depth order, and dump it to fig format.
fig2dev makes an eps file, and then you let the printer take care of it. Filled polygons is something a postscript printer can draw in its sleep; hidden surface removal becomes moot.
Someone already suggested the OpenGl -> PS library, but my suggestion has the benefit of requiring almost no APIs at all; just rotate, decimate, sort, and dump to file.
Johan
bit of my favorite hooby horse, so I hope you'll excuse me repeating it:
ALL key-based encryption is in NP (it is easy to verify a suggested key). So it matters little whether it is based on factorisation, points on elliptic curve, rotor positions, or S boxes.
BTW I do believe that while nobody knows whether FACTOR is NP, it is commonly held NOT to be NP COMPLETE.
Besides, the copper that some chips are made of is a pretty good conductor of heat already. Perhaps all that we need are copper rods embedded in chips (cunningly placed to avoid short circuits, and equally cunningly shaped to avoid inductive effects on the useful copper) to help draw out the heat?
But then I think that if I had actually read the article, I might have something more insightfull to say.
*clap*
*clap*
*clap*
bravo. well said.
Since the originals are of much lower quality than movies, that lowers the bandwidth requirements even more. Are we looking at 150MB for an hour's worth of entertainment?
I personally don't have a television, but I do have sneakernet fast internet access with my laptop, so if I were able to rent/download/subscribe to just-past-their-prime television series to play in the BG while I cook, I'd be all over that.
However, the price has to be right. $10 a month gets me basic cable, and $200 gets me an All-In-Wonder to skip comercials with, so they have to compete with those prices.
I wonder whether bandwidth prices for the studios can make that worth their while.
Thing is, I read your comment before I read the article, and I'm thinking to myself, "Cool, berlin has their display-postscript driver" --which I assumed existed because someone else mentioned NeWS -- "implemented in forth?! Nifty".
I was more than a little bummed to find out I was somewhat off.
It would take a grand total of four apps ported to berlin to make it a viable alternative for me.
1) emacs
2) xdvi/xdpf (they wouldn't be called that of course)
3) xterm
4) netscape (or the frankly superior Internet Explorer. I wish it weren't so, but t'is.)
Those are the only gui apps that I've used during the daylight hours for the last four odd years. Incidentally, they are also the apps that would benefit most from anti-aliasing (preferably sub-pixel).
The article didn't quite make it clear to me whether an X compatiblity layer would be possible, either at the developer level (recompile all apps) or the user level (just start the Xf86 server that hooks up with a running berlin server). Anyone want to chip in with an opinion?
Also, (somewhat OT): does anyone have a status update for Xrender in the newest Xfree86? I seem to recall a page promising sub-pixel anti-aliasing for my laptop, and I want it badly. That would make my emacs SOO easy to read.
The recent release was mum tho, so I'm sorta in the dark.
I'm suprised that this sort of thing isn't better done in orbit, where at least gravity and temparature are on your side. You could imagine slipping a few big mirrors up to the lagrange points, and focussing them on hubble or something for the inferomtery.
Just a thought.
That must be the first time the lameness filter actually filtered junk!
:-P
seriously, that lameness filter has pissed me off so many times (ever spend 15 minutes drawing ascii art to illustrate a point and then
well, the best(*) a 100% parallel system could do in that configuration would be to equal the faster chip. 2 x 500 = 1 x 1000 !
Now if the article had said that a 2 x 800 was just barely as fast as a 1 x 1000, under an OS with the capability to use both CPUs, that would have been noteworthy. But this must have been one of the least informative comparisons I've heard in a long time.
(*) disregarding memory bandwidth, which depends on busses and what not.
hrm.
I was under the impression that we were coming out of a small ice age. This is claimed as evidence for winters being whiter in your grandparent's days.
This is often claimed as the real reason why the earth is warming up, as opposed to greenhouse effects, which are presumed to take longer to gestate. (I must admit that I had thought it strange that we had already had an effect on the climate -- which is not to say that our current emissions won't wreack havoc in the next centuries).
I'm reading the book with Feyman's talks and lectures (some of them) called "the joy of finding stuff out" or something similar, and he also mention the work of eacly 20 century mathematician ... forget the name... Hilbert? Anyways, this guy went off and worked out all these therories about how vectors and matricies that appeared to have no use whatsoever. Until quantuum mechanics came around, and Pauli had to reinvent much of it.
Now Pauli was a smart guy, so a lesser intellect might have looked up Hilbert's stuff rather than reinventing it.
This is an instance of the old ATM vs IP or CISC vs RISC debates. It's the old engineering tradeoff: work smart but slow or work quick and dirty. Tile based rending is an instance of smart and slow, ie they do no more work than they have to, and thus get away with slower clocks and memory. The NVIDIA card is quick and dirty.
Historically, it is almost always the case that quick and dirty is the cheaper way to go, as it allows economies of scale to come into play. However, it is seeming more and more like the memory bandwidth bottleneck is here to stay, so the smart and slow approach is looking pretty good. Likewise as we run into physical limitations for network bandwidth, IP is going to have a harder and harder time to provide acceptable QoS and multicast solutions and ATM-like technologies will start becoming more prevalent.
erm.
que?
napster I thought supported whatever format people saved their songs as. I was under the impression (never tried it, usenet for me) that napster didn't actual create (or until recently) screen any files.