Well, the $3 seemed to be like a "service change": $1->4ILS, food in telavivi ~100(?) -> $25. So for about 1/8th of meal, you can get your kid looked after. Beats a babysiter! Now try charging 5x MAX($babysitter,$own-salary), and they will pickup the rugrats at the speed of light.
Penalties should be proportionnal to resources available. If all the parents make big ILS, then the penalty needs to match up. Pushing the limit way-out on the 1st time will cause them to take the kids somewhere else. Tricky balance... TCP's RTO exponential penalty?
" - Hi, this is the 3rd time you pickup your kid late... we'll have to charge you an arm... please step towards the chopping block...
- haaaaaaarrrgh!
- The doctor will see u now. "
Anybody remember the "GE LightSpeed Ultra CT scanner" ad? The one that can scan a beating heart.
It was a precursor to this 1s ad thing. It would show a series of well known pictures/art. Each on screen for only a few (5) frames (frame == tivo slowmotion frame). The ad gave an impression of something wrong. After catching it on my TiVo, and single-stepping it became apparent: those old black-and-white pictures and egyptian murals where animated. That was cunning, distorting the familiar only very slightly. Somebody had to come up with that idea to catch attention within a few frames.
Just good marketting. I'm old enough to remember the days when Compuserve (before being bought by AOL) had real TCP/IP that allowed Minix users to connect over a modem while AOL had some hacked up custom protocol that worked by replacing all network related apps on windows.
Look who's still around... That's Marketting with a capital M or maybe just lazy/ignorant users.
The new trend is to have something above admin to control DRM and other such sensitive data. So the system can grant itself more priviledges than the Administrator.
OS vendors are now beleiving they can handle the internals better than the admin. And in some cases they are right. Just look at how many "Admins" end up with viruses running as "Admin" because they don't patch or just feel the love from clicking on that unknown EXE attachement. The Mac OS X has that same "admin group" vs. "root user" thing going: root access is disabled by default (see Mac's FAQ on that).
But in the DRM-rich future there will be a need for something above admin. Phoenix BIOS is starting that kind of stuff with its TrustedCore product. Maybe even your laptop has on of those partitions that a even Admin can't get too (but just boot linux, or take the drive to another machine;) ).
What if it is not your site that is being targeted, but just the google ads.
Denial of revenue. If somebody would go to sites with a script that makes click-through based ads not worth the investment, then that site would be banned, and any real revenue it could have generated for google is now gone.
I'm sure google is working very hard not to ban sites that could potentially earn them money.
A PSP is close to what you want. The standby feature is quite handy: need to pay some attention to the significant other during a boss-battle? Push that button, and done... game saved in memory. No more 10x "5 more minutes hunny...". Switching games in the middle is not possible, but now games like GTA or Untold Legends are manageable. I carry my PSP everywhere (baggy pants with 400 pockets) and play in 3min-slices while picking up somebody for carpooling, waiting in line at the supermarket,...
And most probably the data streaming through is repeated at regular intervals (e.g. from the military stuff, vehicle coords, gun angles,...).
So, if you need extra info on some result (e.g. look for out-of-place vehicles, followed by, what are the drivers vitals?) you just run another query on the new data stream and definitively don't do look at past data. That is where the "don't store" comes in.
I'm sure the Patent Office does not have a patent on patenting. It would be very likely (in view of their recent non-efforts to search for prior art) that the patent would be granted. Even better, patent receiving a patent-grant... now that's a money maker.
At work we often use it to see how much of the system is hung.
Pressing capslock and getting the LED on the keboard to change involves a lot of system components: Hit key -> kbd generates packet for SuperIO - SuperIO sends info to SouthBridge over LowPinCount bus (or whatever else the chipset wants) - SouthBridge generates interrupt to CPU over SB/CPU link (AMD+ALI use HyperTransport)
(depending on Interrupt controller configuration, CPU might need to reply to SB to get the interrupt vector) - CPU jumps to code that handles keyboard interrupt
(this involves CPU seeing interrupt signal and reading memory for code) - interrupt-handler/keyboard-driver reads keyboard IO port (single/multiple ports) - CPU request IO from SB - SB requests data from SuperIO - data from superIO goes back same path - int-handler/driver sees CapsLock, toggles states by writing to keyboard IO port - CPU->SB->SuperIO->Keyboard and LED toggles
I work in the CPU business, and when a customer design hangs it is good to know that all of the above is still ok. --
I actually worked on a CAN-protocol CPU participating in mainting car stability and engine control. The scary bit, is that if exposed to high frequency (4x the CPU clocks) the RISC CPU would freeze (meta-stable state, where some bits are not 0 nor 1)! (found this out by accidentally mis-configuring a test-patern generator)
So if your car's power distribution to the wheels was being told to focus on the rear right, and reduce input to front left, while the suspsension was ajusted accordingly and the CPU froze: fly baby! That system shipped in most big german cars.
Luckily, the system had 3 CPUs (16bit DSP, 8bit DSP, 8bit RISC) watching each others backs (if one failed to respond to requests of the others it would get reset, and its recovery time was within the critical limits. ouf!)
Ever heard the phrase "I was driving 25, and my airbag popped!", now you know why.
Any software that can receive remote commands should have a built-in public-key, and only the writers of the software would have the matching private key. The software should check the digital signature of any incomming command, and only execute the authenticated ones. That way, no-body could exploit the remote-command feature.
Check the phoenixnet backdoor at http://slashdot.org/yro/01/06/19/2039216.shtml it makes ES5's remote command scheme look sooo laaaame... PhoenixNet used to have remote controled sw that would only execute digitally signed commands. Over 200000 PC (UUID'd) would phone home every day requesting the latest set of digitally signed commands. Even if somebody DNS-forged the master server's name, they still could not send authenticated commands to those PCs. Other searches for "phoenixnet download" show that the scheme has been shutdown since some time.
It seems like a waste of time. Every month somebody finds on overflow exploit (slight exageration based on the double RPC patch). Even if the source code was checked for backdoors, and then compiled in China, there would still be new exploits every month. A hostile party would only have to study the binary and keep eploits secret until needed.
http://www.aecom.yu.edu/aif/instructions/pictrogra phy/printer1.htm This is a true RBG printer (not with internal rgb -> cymk convertion like some of the Epson 1270)
How about using RGB instead? and the background would be K instead of white. Or maybe double layered thingy, with RBG on top, K, then the white at the bootom. The K cells would be under the RG and B.
Neal Stephenson's "The Diamond Age or, A young Lady's illustrated primer." focuses on a book that uses re-arangeable ink (a book with an AI). Nice reading for SF fans.
from Rafsec's web http://www.rafsec.com/products/pallet_set.htm
"Because Rafsec is a multi-protocol, multi-frequency supplier of RFID transponders, the Wooden Pallet Transponder can be used with any RFID technology, from low-cost read-only to higher-cost encrypted read-write memory."
Say yes to RFIDs, but only if they are disabled after initial use. Passing the doors of the store could tell the RFID to stop responding.
> Naturally, I want to share this information with > their software engineers, and would even be nice > enough and suggest a means to fixing it. Drop the ego! Go to a public library and email the appropriate people anonymously. If you still want to show off your skills to your buddies, then gpg sign your email, so that later you can prove that you wrote it.
Slashdot Mirror
... I'm afraid my kids will read this and grow up to be offended by everything.
Don't expect fast deliveries, they're not UPS. They only do 1.3mm per day (15nm/s). And lets hope direction is not an issue.
--
http://www.med.harvard.edu/publications/On_The_Bra in/Volume3/Number4/Cochlear.html
Upgrades tend to be painfull.
After the StarTak, the Plantronics...
just need the warp-drive.
"financial incentives or penalties"
Well, the $3 seemed to be like a "service change": $1->4ILS, food in telavivi ~100(?) -> $25.
So for about 1/8th of meal, you can get your kid looked after. Beats a babysiter!
Now try charging 5x MAX($babysitter,$own-salary), and they will pickup the rugrats at the speed of light.
Penalties should be proportionnal to resources available. If all the parents make big ILS, then the penalty needs to match up. Pushing the limit way-out on the 1st time will cause them to take the kids somewhere else.
Tricky balance... TCP's RTO exponential penalty?
" - Hi, this is the 3rd time you pickup your kid late... we'll have to charge you an arm... please step towards the chopping block...
- haaaaaaarrrgh!
- The doctor will see u now.
"
The ad ("Images") mentionned above that uses sub-second sequences to catch attention is at the bottom ofi ng/tele_ads.htm
http://www.ge.com/en/company/companyinfo/advertis
Anybody remember the "GE LightSpeed Ultra CT scanner" ad? The one that can scan a beating heart.
...)
It was a precursor to this 1s ad thing.
It would show a series of well known pictures/art. Each on screen for only a few (5) frames (frame == tivo slowmotion frame). The ad gave an impression of something wrong. After catching it on my TiVo, and single-stepping it became apparent: those old black-and-white pictures and egyptian murals where animated.
That was cunning, distorting the familiar only very slightly.
Somebody had to come up with that idea to catch attention within a few frames.
(i could not find that ad
Just good marketting.
I'm old enough to remember the days when Compuserve (before being bought by AOL) had real TCP/IP that allowed Minix users to connect over a modem while AOL had some hacked up custom protocol that worked by replacing all network related apps on windows.
Look who's still around...
That's Marketting with a capital M
or maybe just lazy/ignorant users.
The new trend is to have something above admin to control DRM and other such sensitive data. So the system can grant itself more priviledges than the Administrator.
;) ).
OS vendors are now beleiving they can handle the internals better than the admin. And in some cases they are right. Just look at how many "Admins" end up with viruses running as "Admin" because they don't patch or just feel the love from clicking on that unknown EXE attachement.
The Mac OS X has that same "admin group" vs. "root user" thing going: root access is disabled by default (see Mac's FAQ on that).
But in the DRM-rich future there will be a need for something above admin. Phoenix BIOS is starting that kind of stuff with its TrustedCore product. Maybe even your laptop has on of those partitions that a even Admin can't get too (but just boot linux, or take the drive to another machine
What if it is not your site that is being targeted, but just the google ads.
Denial of revenue.
If somebody would go to sites with a script that makes click-through based ads not worth the investment, then that site would be banned, and any real revenue it could have generated for google is now gone.
I'm sure google is working very hard not to ban sites that could potentially earn them money.
A PSP is close to what you want.
The standby feature is quite handy: need to pay some attention to the significant other during a boss-battle? Push that button, and done... game saved in memory. No more 10x "5 more minutes hunny...".
Switching games in the middle is not possible, but now games like GTA or Untold Legends are manageable.
I carry my PSP everywhere (baggy pants with 400 pockets) and play in 3min-slices while picking up somebody for carpooling, waiting in line at the supermarket,...
The human factor is often the weakest link: he got bought by somebody who does not want privacy.
(just being paranoid)
And most probably the data streaming through is repeated at regular intervals (e.g. from the military stuff, vehicle coords, gun angles,...).
So, if you need extra info on some result (e.g. look for out-of-place vehicles, followed by, what are the drivers vitals?) you just run another query on the new data stream and definitively don't do look at past data. That is where the "don't store" comes in.
Patent "being granted a patent", you get to sue more people.
I'm sure the Patent Office does not have a patent on patenting.
It would be very likely (in view of their recent non-efforts to search for prior art) that the patent would be granted.
Even better, patent receiving a patent-grant... now that's a money maker.
At work we often use it to see how much of the system is hung.
Pressing capslock and getting the LED on the keboard to change involves a lot of system components:
Hit key -> kbd generates packet for SuperIO
- SuperIO sends info to SouthBridge over LowPinCount bus (or whatever else the chipset wants)
- SouthBridge generates interrupt to CPU over SB/CPU link (AMD+ALI use HyperTransport)
(depending on Interrupt controller configuration, CPU might need to reply to SB to get the interrupt vector)
- CPU jumps to code that handles keyboard interrupt
(this involves CPU seeing interrupt signal and reading memory for code)
- interrupt-handler/keyboard-driver reads keyboard IO port (single/multiple ports)
- CPU request IO from SB
- SB requests data from SuperIO
- data from superIO goes back same path
- int-handler/driver sees CapsLock, toggles states by writing to keyboard IO port
- CPU->SB->SuperIO->Keyboard and LED toggles
I work in the CPU business, and when a customer design hangs it is good to know that all of the above is still ok.
--
I actually worked on a CAN-protocol CPU participating in mainting car stability and engine control.
The scary bit, is that if exposed to high frequency (4x the CPU clocks) the RISC CPU would freeze (meta-stable state, where some bits are not 0 nor 1)! (found this out by accidentally mis-configuring a test-patern generator)
So if your car's power distribution to the wheels was being told to focus on the rear right, and reduce input to front left, while the suspsension was ajusted accordingly and the CPU froze: fly baby!
That system shipped in most big german cars.
Luckily, the system had 3 CPUs (16bit DSP, 8bit DSP, 8bit RISC) watching each others backs (if one failed to respond to requests of the others it would get reset, and its recovery time was within the critical limits. ouf!)
Ever heard the phrase "I was driving 25, and my airbag popped!", now you know why.
Any software that can receive remote commands should have a built-in public-key, and only the writers of the software would have the matching private key.
t makes ES5's remote command scheme look sooo laaaame...
The software should check the digital signature of any incomming command, and only execute the authenticated ones.
That way, no-body could exploit the remote-command feature.
Check the phoenixnet backdoor at http://slashdot.org/yro/01/06/19/2039216.shtml
i
PhoenixNet used to have remote controled sw that would only execute digitally signed commands.
Over 200000 PC (UUID'd) would phone home every day requesting the latest set of digitally signed commands. Even if somebody DNS-forged the master server's name, they still could not send authenticated commands to those PCs.
Other searches for "phoenixnet download" show that the scheme has been shutdown since some time.
--
jpa
It seems like a waste of time.
Every month somebody finds on overflow exploit (slight exageration based on the double RPC patch).
Even if the source code was checked for backdoors, and then compiled in China, there would still be new exploits every month.
A hostile party would only have to study the binary and keep eploits secret until needed.
--
jpa
my bad. Ignore the previous post, that printer still does cym internally. Damn. I'm sure RGB could work, but would probably look ugly...
http://www.aecom.yu.edu/aif/instructions/pictrogra phy/printer1.htm
This is a true RBG printer (not with internal rgb -> cymk convertion like some of the Epson 1270)
So RGB/K/W could work just fine.
How about using RGB instead? and the background would be K instead of white. Or maybe double layered thingy, with RBG on top, K, then the white at the bootom. The K cells would be under the RG and B.
Neal Stephenson's "The Diamond Age or, A young Lady's illustrated primer." focuses on a book that uses re-arangeable ink (a book with an AI).
Nice reading for SF fans.
from Rafsec's web http://www.rafsec.com/products/pallet_set.htm
"Because Rafsec is a multi-protocol, multi-frequency supplier of RFID transponders, the Wooden Pallet Transponder can be used with any RFID technology, from low-cost read-only to higher-cost encrypted read-write memory."
Say yes to RFIDs, but only if they are disabled after initial use. Passing the doors of the store could tell the RFID to stop responding.
> Naturally, I want to share this information with
> their software engineers, and would even be nice
> enough and suggest a means to fixing it.
Drop the ego!
Go to a public library and email the appropriate people anonymously.
If you still want to show off your skills to your buddies, then gpg sign your email, so that later you can prove that you wrote it.
--
jpa