Apparently @Home is looking for the little bit of extra revenue they can get by selling additional IPs to people (like me) who have more than one computer.
(1)If you're computers were at different location, you'd need multiple connections (and multipe ip's anyway), which is what VPN's are used for, securing communication between remote points.
(2) If you have more than one computer at your home -- i'm assuming you can still get a netopia router that will dynamically assign all your computers with local (non-routable) ip addresses and then act as a kind of IP-masq, so you can still have more than one computer on you're local area network _or_
(3) if you have a persistant connection, you should probably configure a box a a firewall anyway, using something like ip-chains (*NIX) or Proxy Server (NT) [if you use ip-chains you'll need to use a separerate applications proxy - i recommend squit.
Anyways -- the point is, this will have no effect on having more than one coputer at the same location.
To all of those who are posting the 'one more reason to use encryption' posts, do you honestly think that big brother won't just set up they're box to save and store all encrypted communication? or add the sender and reciever to a special 'potential trouble' list. And yes, they can tell if it's encrypted, because encryption, or at least good encryption, does obey a certain statistical pattern (i.e. plaintext will be have a high percentage of recurring character, while ciphertext should be totally random). Granted, compression does something simialer, but still -- I'm on enough lists as it is!
I found this quote on cnet's article about the aclu's objection especially telling "Carnivore is roughly equivalent to a wiretap capable of accessing the contents of the conversations of all of the phone company's customers, with the 'assurance' that the FBI will record only conversations of the specified target," read the letter. "This 'trust us, we are the government' approach is the antithesis of the procedures required under our wiretapping laws."
It's about time! Russia has already held up the launch, what, three times now? This is a bit surprising, though, since I just read at canoe that the date was expected to be announced tomorrow. I want to know how well this is going to be inspected before launch. Do we really want the ISS failing because somebody didn't want to save face? (Let's hope ISS works better that IIS 8p ).
If you *have* to run some 'win32' only apps (and I'm assuming you have a few, because you get a notebook), take a look at wine (http:\\www.winehq.com). They've come a long way! they can even support native windows DLL's which does wonders for their compatibility, and it support's my only 'win32 only' app. (starcraft). if it works out for you, you even get to take your notebook home =)
This is correct -- the majority of the DNA is junk, but the junk can easily be parsed out, because of a sequence that works essentially like a stop-bit'. In actualy gene's, these only occur at the end of a sequence, which could be hundreds or thousands of pairs long, but in the junk area, they have a tendancy to accumulate. Therefore, when you have three or four of these stop-bits in a sequence every 20 or so pairs, you can through that area out as junk.
As far as actually knowing what a gene does without flipping it and seeing what it does - unfortunately, are understanding of the way genetics actually works is still mostly from black-box engineering. In fact, we still don't have the ability to predict how a giver protien will fold in three dimensions!
but what i want to know is what kind of database these guys are using?
As somebody has already pointed out, nueral nets are old hat in AI. And not all of them are computational models of neural nets, some are actually embedded. I've used a simple neural net in a stiqueto-based robot to control it's leg movement. The only way a hardware based model would be new is if it actually emulated the brains electro-chemical reactions instead of using digital logic. The problem with hardware implementations (as opposed to computer modeling is: Every neuron has axon's connected to other neurons. This is called Fan-out. In the human mind, average fan-out is somewhere along the lines of 1000, as obossed to an artifical net, which has a avereage fan-out of 10. The problem with implementing this in hardware, even with a fanout as little as 10, is attenuation. Every time the neuron fires, the signal looses strenght. Because real nuerons are electro-chemical, and are self-powered. they don't have this problem. In a small net (like in my stiqueto-bot), i just used a couple a 'strategically' placed op-amps (my strategy involved poking it with a meter =). But in a net with a fan-out of 1000. This wouldn't be practical.
Where I work, we have an online store, and we also have a *real* store up infront of our offices. We do a lot of mail- and phone- order to. (our clients are realtors, so they're always ordering real-estate forms and what-not, but they're not to tech-savy). I have our website set up so whenever anybody orders something online, it sends out three emails: a confirmation to the person (who still clicks 'submit' five or six time -got i wish i could kill those idiots), one to the retail clerks in the 'real' store, and one to shipping. The clerks in the real store enter the credit card information manually in the same system they use to bill normal credit cards, and the shipping guy ships it. This system isn't perfect, as it still has double entery and wouldn't work for a high-sales sight, but it'll work for any low-volume sight that is already set up to take credit cards other ways. I'm not sure this'll help whoever sent in the question, but hopefully someone who is reading this might be able to use this model. If you already have a website that's 'information only', i.e. your corprate broshure scanned onto the internet, it's a great idea to take to the suits 8p
As somebody has already pointed out, nueral nets are old hat in AI. And not all of them are computational models of neural nets, some are actually embedded. I've used a simple neural net in a stiqueto-based robot to control it's leg movement. The only way a hardware based model would be new is if it actually emulated the brains electro-chemical reactions instead of using digital logic. The problem with hardware implementations (as opposed to computer modeling is:
Every neuron has axon's connected to other neurons. This is called Fan-out. In the human mind, average fan-out is somewhere along the lines of 1000, as obossed to an artifical net, which has a avereage fan-out of 10. The problem with implementing this in hardware, even with a fanout as little as 10, is attenuation. Every time the neuron fires, the signal looses strenght. Because real nuerons are electro-chemical, and are self-powered. they don't have this problem. In a small net (like in my stiqueto-bot), i just used a couple a 'strategically' placed op-amps (my strategy involved poking it with a meter =). But in a net with a fan-out of 1000. This wouldn't be practical.
Well...C++ doesn't allow assignment to a NULL pointer, so void * NULL; NULL = (void*)0
technically shouldn't work, even though it does on _every_ c compiler I've seen, inless ANSI_STRICT or some such is turned on.
great, let's do what EVERY other vendor on the planet does, and say it's a brand new innovation! we invented a new language! but seriously, I'm sure they have enough *value added* features to make sure it's not compatible with C/C++. elements of java probably means that class definition and decleration will be combined.
Most designs in nanotechnology I've seen (most notably a lot of Eric Drexler's structures) are built out of carbon atoms. It my understanding of this is right, then structures made of carbon atoms have the potential to bond under preasure, and they'd definately be under pressure in mechanical systems. I'm still a nano-fan though -- because nan* is cool =)
This point is absolutely true. I work in the MIS department of a relatively small company ( Recent documents. when i see this a take heart, though. It means microsoft is not nearly entrenched as people seem to think...only the GUI 8p.
Who cares if the g4's clockspeed isn't as high? The fact that everybody is so opssessed with their clock speeds is a result of intel's indoctrinization of 'clock speed good- archetectual advanced bad'. The mac g4 can't be exported because it's legally classified as a supercomputer, and therefore has 'stratigic military value'. It has a 128-bit RISC processor, while intel's still living in 32-bit CISC land (but they're comming out with the IA-64 soon. really. they said so =).
Mac's are also SCSII internally, which increases disk performance, etc., etc.. The only reason not to get an iMac is because they don't come in black...
linux wasn't meant to be an 'enterprise' operating system. It was written as an experiment on the memory management capabilities of the 386. The reason linux is getting so much recognition in the b2b world and in ecommerce is because linux stands on it's own merits. That's why we all use it, not because of some marketing hype. Okay...linux distributions start changing there name's to appease the suits. what next? 'value' added features, fragmentation of the distro's, focusing on marketing insead of programming...where does it end? no...slackware should keep it's name, and be judged on it's merit as an operating system (well...distribution). If the suits don't think 'slackware' if professional enough let them use NT. But we'll all laugh at them when their server goes down.
True.... What we need are encrypted core dumps encrpted swap is nice...but the big problem with memory is the core dump. If someone has local access to a box, and can get a core dump, there's a change they can get login-name/password-hash combo's, and it's trivial to run a word-list through a program like crack. on improperly configured boxes, local access isn't even required as long as their running apache, because apache let's you enter the '..' directory in path names.
hacking sites archive (i recomend attrition.org. Go to the defaced archive, and it'll tell you what operating system the hacked b0x was running. The vast majority of the hacking sites are running nt. Windows NT has security holes --they're just not as well documented. Security through obscurity. [is your friendly neighborhood hackers installed there software as an NT service, you might not have noticed.]
radius - us = radii ...um....it doesn't exactly work like that. Changing the sufix derives from latin (virii is second declension. The full lexicon for would be: virus, -i, m.,
Porting microsoft applications to other platforms wouldn't be that easy. It would probably entail a substantial amount of code-rewritting, -either that, or porting all the back-end stuff to other OSes, and users of other OSes don't wan't all the bloat that that would entail. Well, there's WINE whcih might make things a bit easier, but, say, what about COM? are they going to use GNOME? DCOM? how's that going to implemented on other platfroms? And DirectX relies on COM, so no COM = no Direct X. and Active X? I suppose they could port that, but I'll just use a widget, thank you very much. Microsoft burned themselves with their integration. Porting most MS programs that take full advantage of their 'integration' would mean either porting a substantial part of their platform, or substiting with native services, and I doubt man of us Linux users are going want to install half the bloat that comes with windows so we can run word when we have Emacs and vi. but to play Final Fantasy VIII...hmmmmmm....
anybody who's ever done any work with robotics knows that the biggest problem involved in the construction of autonomous mobile robots is finding a power supply that's not to heavy. In most robots, the power suppy makes up a substantial part of the mass (which is why we probably won't have gecko shoes any time soon. Yeah, the technologies cool, but right now, what we need in robotics is a breakthrough in power-supplies. Either that or we could run them on Transmuta's new chip (with S.u.S.E) 8p
Finally. I was a Borland zealot back in the dos days, but there C compilerer just didn't make it into the 32-bit world. Maybe now we'll get a free online help system instead of having to pay for the damn msdn.
Actually, Earthlink-Mindspring is now America's Largest ISP. And Netscape 6 is based on Mpzilla. (The prieview release is basically milestone 14 with a different splash screen). If they ever get arround to implementing layers, it'll be a pretty nice browser.
Slashdot Mirror
Apparently @Home is looking for the little bit of extra revenue they can get by selling additional IPs to people (like me) who have more than one computer. (1)If you're computers were at different location, you'd need multiple connections (and multipe ip's anyway), which is what VPN's are used for, securing communication between remote points. (2) If you have more than one computer at your home -- i'm assuming you can still get a netopia router that will dynamically assign all your computers with local (non-routable) ip addresses and then act as a kind of IP-masq, so you can still have more than one computer on you're local area network _or_ (3) if you have a persistant connection, you should probably configure a box a a firewall anyway, using something like ip-chains (*NIX) or Proxy Server (NT) [if you use ip-chains you'll need to use a separerate applications proxy - i recommend squit. Anyways -- the point is, this will have no effect on having more than one coputer at the same location.
"Well it seems the MPAA (Motion Picture Association of America) was feeling left out of all the suing action going on "
the mpaa isn't feeling left out of anything, they've been suing 2600 for awhile now.
To all of those who are posting the 'one more reason to use encryption' posts, do you honestly think that big brother won't just set up they're box to save and store all encrypted communication? or add the sender and reciever to a special 'potential trouble' list. And yes, they can tell if it's encrypted, because encryption, or at least good encryption, does obey a certain statistical pattern (i.e. plaintext will be have a high percentage of recurring character, while ciphertext should be totally random). Granted, compression does something simialer, but still -- I'm on enough lists as it is!
I found this quote on cnet's article about the aclu's objection especially telling "Carnivore is roughly equivalent to a wiretap capable of accessing the contents of the conversations of all of the phone company's customers, with the 'assurance' that the FBI will record only conversations of the specified target," read the letter. "This 'trust us, we are the government' approach is the antithesis of the procedures required under our wiretapping laws."
It's about time! Russia has already held up the launch, what, three times now? This is a bit surprising, though, since I just read at canoe that the date was expected to be announced tomorrow.
I want to know how well this is going to be inspected before launch. Do we really want the ISS failing because somebody didn't want to save face? (Let's hope ISS works better that IIS 8p ).
If you *have* to run some 'win32' only apps (and I'm assuming you have a few, because you get a notebook), take a look at wine (http:\\www.winehq.com). They've come a long way! they can even support native windows DLL's which does wonders for their compatibility, and it support's my only 'win32 only' app. (starcraft).
if it works out for you, you even get to take your notebook home =)
It's Bill's fault! if this happened on my operating system, i could just zcat the programmers man page! (don't need no fansy 'man' commond =).
This is correct -- the majority of the DNA is junk, but the junk can easily be parsed out, because of a sequence that works essentially like a stop-bit'. In actualy gene's, these only occur at the end of a sequence, which could be hundreds or thousands of pairs long, but in the junk area, they have a tendancy to accumulate. Therefore, when you have three or four of these stop-bits in a sequence every 20 or so pairs, you can through that area out as junk.
As far as actually knowing what a gene does without flipping it and seeing what it does - unfortunately, are understanding of the way genetics actually works is still mostly from black-box engineering. In fact, we still don't have the ability to predict how a giver protien will fold in three dimensions!
but what i want to know is what kind of database these guys are using?
As somebody has already pointed out, nueral nets are old hat in AI. And not all of them are computational models of neural nets, some are actually embedded. I've used a simple neural net in a stiqueto-based robot to control it's leg movement. The only way a hardware based model would be new is if it actually emulated the brains electro-chemical reactions instead of using digital logic. The problem with hardware implementations (as opposed to computer modeling is: Every neuron has axon's connected to other neurons. This is called Fan-out. In the human mind, average fan-out is somewhere along the lines of 1000, as obossed to an artifical net, which has a avereage fan-out of 10. The problem with implementing this in hardware, even with a fanout as little as 10, is attenuation. Every time the neuron fires, the signal looses strenght. Because real nuerons are electro-chemical, and are self-powered. they don't have this problem. In a small net (like in my stiqueto-bot), i just used a couple a 'strategically' placed op-amps (my strategy involved poking it with a meter =). But in a net with a fan-out of 1000. This wouldn't be practical.
Where I work, we have an online store, and we also have a *real* store up infront of our offices. We do a lot of mail- and phone- order to. (our clients are realtors, so they're always ordering real-estate forms and what-not, but they're not to tech-savy). I have our website set up so whenever anybody orders something online, it sends out three emails: a confirmation to the person (who still clicks 'submit' five or six time -got i wish i could kill those idiots), one to the retail clerks in the 'real' store, and one to shipping. The clerks in the real store enter the credit card information manually in the same system they use to bill normal credit cards, and the shipping guy ships it. This system isn't perfect, as it still has double entery and wouldn't work for a high-sales sight, but it'll work for any low-volume sight that is already set up to take credit cards other ways.
I'm not sure this'll help whoever sent in the question, but hopefully someone who is reading this might be able to use this model. If you already have a website that's 'information only', i.e. your corprate broshure scanned onto the internet, it's a great idea to take to the suits 8p
As somebody has already pointed out, nueral nets are old hat in AI. And not all of them are computational models of neural nets, some are actually embedded. I've used a simple neural net in a stiqueto-based robot to control it's leg movement. The only way a hardware based model would be new is if it actually emulated the brains electro-chemical reactions instead of using digital logic. The problem with hardware implementations (as opposed to computer modeling is:
Every neuron has axon's connected to other neurons. This is called Fan-out. In the human mind, average fan-out is somewhere along the lines of 1000, as obossed to an artifical net, which has a avereage fan-out of 10. The problem with implementing this in hardware, even with a fanout as little as 10, is attenuation. Every time the neuron fires, the signal looses strenght. Because real nuerons are electro-chemical, and are self-powered. they don't have this problem. In a small net (like in my stiqueto-bot), i just used a couple a 'strategically' placed op-amps (my strategy involved poking it with a meter =). But in a net with a fan-out of 1000. This wouldn't be practical.
Well...C++ doesn't allow assignment to a NULL pointer, so
void * NULL;
NULL = (void*)0
technically shouldn't work, even though it does on _every_ c compiler I've seen, inless ANSI_STRICT or some such is turned on.
great, let's do what EVERY other vendor on the planet does, and say it's a brand new innovation! we invented a new language!
but seriously, I'm sure they have enough *value added* features to make sure it's not compatible with C/C++.
elements of java probably means that class definition and decleration will be combined.
Most designs in nanotechnology I've seen (most notably a lot of Eric Drexler's structures) are built out of carbon atoms. It my understanding of this is right, then structures made of carbon atoms have the potential to bond under preasure, and they'd definately be under pressure in mechanical systems.
I'm still a nano-fan though -- because nan* is cool =)
This point is absolutely true. I work in the MIS department of a relatively small company ( Recent documents.
when i see this a take heart, though. It means microsoft is not nearly entrenched as people seem to think...only the GUI 8p.
Who cares if the g4's clockspeed isn't as high? The fact that everybody is so opssessed with their clock speeds is a result of intel's indoctrinization of 'clock speed good- archetectual advanced bad'. The mac g4 can't be exported because it's legally classified as a supercomputer, and therefore has 'stratigic military value'. It has a 128-bit RISC processor, while intel's still living in 32-bit CISC land (but they're comming out with the IA-64 soon. really. they said so =).
Mac's are also SCSII internally, which increases disk performance, etc., etc..
The only reason not to get an iMac is because they don't come in black...
linux wasn't meant to be an 'enterprise' operating system. It was written as an experiment on the memory management capabilities of the 386.
The reason linux is getting so much recognition in the b2b world and in ecommerce is because linux stands on it's own merits. That's why we all use it, not because of some marketing hype.
Okay...linux distributions start changing there name's to appease the suits. what next? 'value' added features, fragmentation of the distro's, focusing on marketing insead of programming...where does it end?
no...slackware should keep it's name, and be judged on it's merit as an operating system (well...distribution). If the suits don't think 'slackware' if professional enough let them use NT. But we'll all laugh at them when their server goes down.
True.... What we need are encrypted core dumps
encrpted swap is nice...but the big problem with memory is the core dump. If someone has local access to a box, and can get a core dump, there's a change they can get login-name/password-hash combo's, and it's trivial to run a word-list through a program like crack.
on improperly configured boxes, local access isn't even required as long as their running apache, because apache let's you enter the '..' directory in path names.
- any
hacking sites archive (i recomend attrition.org. Go to the defaced archive, and it'll tell you what operating system the hacked b0x was running. The vast majority of the hacking sites are running nt. Windows NT has security holes --they're just not as well documented. Security through obscurity. [is your friendly neighborhood hackers installed there software as an NT service, you might not have noticed.]radius - us = radii
...um....it doesn't exactly work like that. Changing the sufix derives from latin (virii is second declension. The full lexicon for would be: virus, -i, m.,
Johnny Mnomonic had what? 100 Mb? Something Like that?
Thanks....but I'll just stick a zip-disk in my pocked for now.
well... my linux distro didn't come with solataire, and rebooting is kina a pane....
Porting microsoft applications to other platforms wouldn't be that easy. It would probably entail a substantial amount of code-rewritting, -either that, or porting all the back-end stuff to other OSes, and users of other OSes don't wan't all the bloat that that would entail.
Well, there's WINE whcih might make things a bit easier, but, say, what about COM? are they going to use GNOME? DCOM? how's that going to implemented on other platfroms? And DirectX relies on COM, so no COM = no Direct X. and Active X? I suppose they could port that, but I'll just use a widget, thank you very much.
Microsoft burned themselves with their integration. Porting most MS programs that take full advantage of their 'integration' would mean either porting a substantial part of their platform, or substiting with native services, and I doubt man of us Linux users are going want to install half the bloat that comes with windows so we can run word when we have Emacs and vi.
but to play Final Fantasy VIII...hmmmmmm....
anybody who's ever done any work with robotics knows that the biggest problem involved in the construction of autonomous mobile robots is finding a power supply that's not to heavy. In most robots, the power suppy makes up a substantial part of the mass (which is why we probably won't have gecko shoes any time soon. Yeah, the technologies cool, but right now, what we need in robotics is a breakthrough in power-supplies. Either that or we could run them on Transmuta's new chip (with S.u.S.E) 8p
hmmm..........I live about three miles away from my work, and we have three t1 lines here which nobody uses after five...i think i have an idea
Finally. I was a Borland zealot back in the dos days, but there C compilerer just didn't make it into the 32-bit world. Maybe now we'll get a free online help system instead of having to pay for the damn msdn.
Actually, Earthlink-Mindspring is now America's Largest ISP. And Netscape 6 is based on Mpzilla. (The prieview release is basically milestone 14 with a different splash screen). If they ever get arround to implementing layers, it'll be a pretty nice browser.