I didn't know they had a pop-up blocker - thanks, I learned something! Just whenever I use AOL, I'm amazed by how many windows it opens, but I guess that's content and mostly not ads.
AOL has used cached proxies for a while; they've got so many customers, this is worthwhile. If 10 customers per second access cnn.com, then might as well cache it and retreive it only every 30 seconds.
For graphics, AOL bought technology to compress beyond JPEG. It's called ART and I think it works with wavelets. It's cool, but not an industry standard. They purposefully set it for less image quality, so I think that's where the majority of the compression comes from (and not the super-efficency of the algorithm).
Yep, your file will go through at 0.9x of the regular speed (slower). This is less than 5x faster, so they win! All they are guaranteeing is a maximum speed (5x faster), and that's not hard to do. Stupid, yes. Truth in advertising, yes.
The vast majority of 56k modems already do compression, CSLIP compresses headers, and HTML compression is already built into modern browsers. What's left is caching, image-size/quality reduction, and pop-up blocking. AOL already does two of those three - take a guess which two!!
Thanks for the answer. I guess the judge fell for it -- he said "The damage caused by large amounts of e-mail not reaching their destinations should be covered by the sender", but what he should have said was "The loss of potential revenue should be covered by the sender", which is what actually happened. Why should NTT expect to make a profit on illegitimate messages?
I like the verdict and think that the fine is appropriate, but I don't like how it was calculated. Maybe the article misrepresented it, but charging $0.01 per spam seems excessive.
The article says 880 million undeliverable emails are sent every day. At a penny a piece, that's USD$8.8million / day, or $3.2 billion/year. The company does $42 billion in sales per year, I doubt that they spend 7.6% of their income on spam. Or, for that matter, give me $3b/yr and I'll provide the equipment to totally filter all of their undeliverable mail -- they'll save their shareholders $200 million!.
I just wish they said "it cost us 1 man-year of work to stop this guy" and cost it that way instead of making up numbers per message. It's this kind of unjustified damage estimate that "cost" sun $80 million of money that was good enough to tell a judge under oath, but too bogus to tell their shareholders. A doubt NTT has a $3.2b line-item on their annual report.
(and, as others have pointed out, this 880milMsg/day is misaddressed mail - trivial to filter out and it never consume any expensive RF bandwidth)
What copyright you ask? The microcode, of course. That line of thinking is working so far for lexmark. But, we don't really need the DMCA to make crystals illegal. You can also get 34 (+3.7%), 36 (+10%), and 38 (+16%), and 44.1kHz (+35%) crystals easily...
A much better solution for intel would be to have a ring oscillator on the chip and compare the input frequency to that... but that would defeat the whole purpose -- you'd be limited to the actual rating of the chip (because faster chips would have faster ring oscillators), not the rating intel sells the chip at.
No, you can debug it, if you want. The web text refers to an old beta, but you can get to the latest beta version by navigating FTP. Assuming, of course, you can get in.
It has a tiny picture of a hand holding the flashlight, labeled "actual size". Do they mean that it's actually the ~1/2 inch size I see on my screen, or does it mean that they aren't using a stunt double super mega-large hand to illustrate the size?
PWM is ok, but you need to get the pulse rate up pretty high, especially since LEDs are so high-bandwidth. The last thing I want on my desk is a bulb with any noticible flicker. To get the rate up high, you need either dedicated hardware (fine if your uC supports it), or you need to waste lots of CPU power. With the RJ45 jack server, I'm not so sure that the operating system would provide enough control - when it talks to the web, will it stop the user program? Does it ever mask interrupts for too long? Does it support user interrupts? Does it have a user-accessible hardware timer? The PIC is great, but is it really going to be cheaper and easier than a dedicated D/A?
For a large-scale production, PWM would be nice (provided you used the proper filtering capacitor, which could cost more than an ADC if the load is high). But for small one-off projects, go ahead and waste the $4 on the chip; it'll save a lot of debugging time.
There are lots of these serial relay drivers... google for it. They have all sorts of nice features (current limit, fault detection, cascadability and are controlable through the parallel port (you have to bitbang the data and clock bits). The webserver above has 3 general purpose I/Os - enough to control a relay driver.
But, you probablly want an actual A/D converter (preferably with a current output) or a digital potentiometer. There are lots of mfgs of these products, but Maxim is pretty liberal with samples (plus they have some neat innovative products!)
Here's one of the first images done in 1926. Of course, this weather image didn't come from a satellite, but they've been doing this stuff for a long time!
Actually, this is 300k baud - much faster than the//c can handle. If you wrote the code in assembly (running with a 1 MHz clock and 4 cycles per instruction), the tightest code ("LDA $C1FF; LDA $C1FF; etc.") wouldn't keep up -- and that doesn't even do anything with the data! Ah, nostalgia. I remember how the//c couldn't scroll the screen reliably at 2400 baud - it's amazing how far things have come, especially when I DMA stuff at a gigabit/sec to/from a RAID.
Thanks, Xerithane. My submission was in the queue for the longest time before it was rejected (I was keeping an eye on it because I realized I had misspelled benetton in the title). When it finally got rejected, I put in in my journal. It's still listed as rejected; Too bad people looking at my user profile will only see my journal entry and not the link to the larger discussion...
Way too complicated for 99% of the people!! It would be much better to invest a little more smarts in the infrastructure considering the complete lack of tech support that'll be available at the stores. Consider this scenario:
Everything is DHCP-friendly, so people don't have to change their network settings. However, all things go through a proxy computer - when it detects a new MAC address, it will return the sign-in webpage for any page you request. Until you enter the PIN from your receipt or a credit-card number, that will be the only page you'll get. (Or, as a bonus, maybe they'll let you browse a couple of other specific pages for free.)
Heck, you could also have the proxy spoof the email, too. When your computer thinks its logging into your corporate mailserver, it's really just logging into the macdonalds computer... and, surprise, no matter what name/password you use, you'll have some new mail... with a link to the mcdonalds sign-in page, of course!
Transportation- How do you think the food gets to your deli or bakery? In the old days, your choice of husband/wife was limited to about a half-day's walk from your home - if you lived in the boonies, then it's tough luck.
Neat story about the well & socializing, thanks!
Regular jobs are still in the majority; consultants are possible because their networking opportunites are greatly expanded with the use of a phone, a car, and sometimes a computer. My contracter friends swear by payroll services, again a "necessary" worksaver.
Like I said, some people live without many of these things; for others they find them far more beneficial than their costs.
Where I used to work, we sold gigabit fibre channel interfaces that ran at about the same speed. I was thinking of how to show off the technology at tradeshows -- we ultimately went with a 2048x1024ish video transfer at high frame rates, but I thought of having CD pass by (like on a carasol[sp]) at the rate their information was played at - 850X sounds much more impressive than gigabit. But, that wouldn't be eye-catching enough, so I thought of 3.5" floppies... but 85 floppies per second was likely to injure someone!!
Slashdot Mirror
I was expecting to at least see some sort of sad/broken/dead/flaming server icon, but the message was just plain text.
:(
Oh well, I'll just have to make my own
I didn't know they had a pop-up blocker - thanks, I learned something! Just whenever I use AOL, I'm amazed by how many windows it opens, but I guess that's content and mostly not ads.
AOL has used cached proxies for a while; they've got so many customers, this is worthwhile. If 10 customers per second access cnn.com, then might as well cache it and retreive it only every 30 seconds.
For graphics, AOL bought technology to compress beyond JPEG. It's called ART and I think it works with wavelets. It's cool, but not an industry standard. They purposefully set it for less image quality, so I think that's where the majority of the compression comes from (and not the super-efficency of the algorithm).
Yep, your file will go through at 0.9x of the regular speed (slower). This is less than 5x faster, so they win! All they are guaranteeing is a maximum speed (5x faster), and that's not hard to do. Stupid, yes. Truth in advertising, yes.
The vast majority of 56k modems already do compression, CSLIP compresses headers, and HTML compression is already built into modern browsers. What's left is caching, image-size/quality reduction, and pop-up blocking. AOL already does two of those three - take a guess which two!!
Thanks for the answer. I guess the judge fell for it -- he said "The damage caused by large amounts of e-mail not reaching their destinations should be covered by the sender", but what he should have said was "The loss of potential revenue should be covered by the sender", which is what actually happened. Why should NTT expect to make a profit on illegitimate messages?
I like the verdict and think that the fine is appropriate, but I don't like how it was calculated. Maybe the article misrepresented it, but charging $0.01 per spam seems excessive.
The article says 880 million undeliverable emails are sent every day. At a penny a piece, that's USD$8.8million / day, or $3.2 billion/year. The company does $42 billion in sales per year, I doubt that they spend 7.6% of their income on spam. Or, for that matter, give me $3b/yr and I'll provide the equipment to totally filter all of their undeliverable mail -- they'll save their shareholders $200 million!.
I just wish they said "it cost us 1 man-year of work to stop this guy" and cost it that way instead of making up numbers per message. It's this kind of unjustified damage estimate that "cost" sun $80 million of money that was good enough to tell a judge under oath, but too bogus to tell their shareholders. A doubt NTT has a $3.2b line-item on their annual report.
(and, as others have pointed out, this 880milMsg/day is misaddressed mail - trivial to filter out and it never consume any expensive RF bandwidth)
Nope. It's still a public performance.
If it uses a 32kHz crystal (like mentioned in the patents), then these circumvention devices wouldn't be allowed under the DMCA:
38.4 kHz crystal (17% faster)
40 kHz crystal (22% faster)
What copyright you ask? The microcode, of course. That line of thinking is working so far for lexmark. But, we don't really need the DMCA to make crystals illegal. You can also get 34 (+3.7%), 36 (+10%), and 38 (+16%), and 44.1kHz (+35%) crystals easily...
A much better solution for intel would be to have a ring oscillator on the chip and compare the input frequency to that... but that would defeat the whole purpose -- you'd be limited to the actual rating of the chip (because faster chips would have faster ring oscillators), not the rating intel sells the chip at.
No, you can debug it, if you want. The web text refers to an old beta, but you can get to the latest beta version by navigating FTP. Assuming, of course, you can get in.
Let's hope it doesn't cocelot.
EETimes has a teardown of the hiptop... Engineers will get a chuckle out of finding the part that won't die!! There's also a block diagram.
(hint: it's a national part on the same side as the processor)
I love the linked image!!
It has a tiny picture of a hand holding the flashlight, labeled "actual size". Do they mean that it's actually the ~1/2 inch size I see on my screen, or does it mean that they aren't using a stunt double super mega-large hand to illustrate the size?
Now, Gcc on a PCI card is something I'd pay for...
PWM is ok, but you need to get the pulse rate up pretty high, especially since LEDs are so high-bandwidth. The last thing I want on my desk is a bulb with any noticible flicker. To get the rate up high, you need either dedicated hardware (fine if your uC supports it), or you need to waste lots of CPU power. With the RJ45 jack server, I'm not so sure that the operating system would provide enough control - when it talks to the web, will it stop the user program? Does it ever mask interrupts for too long? Does it support user interrupts? Does it have a user-accessible hardware timer? The PIC is great, but is it really going to be cheaper and easier than a dedicated D/A?
For a large-scale production, PWM would be nice (provided you used the proper filtering capacitor, which could cost more than an ADC if the load is high). But for small one-off projects, go ahead and waste the $4 on the chip; it'll save a lot of debugging time.
Hope you get Sabrina back soon!
... I meant D/A converter (but I got the link right)
Quick answer: combine RJ45 web server with serial relay driver and presto!
... google for it. They have all sorts of nice features (current limit, fault detection, cascadability and are controlable through the parallel port (you have to bitbang the data and clock bits). The webserver above has 3 general purpose I/Os - enough to control a relay driver.
There are lots of these serial relay drivers
But, you probablly want an actual A/D converter (preferably with a current output) or a digital potentiometer. There are lots of mfgs of these products, but Maxim is pretty liberal with samples (plus they have some neat innovative products!)
What... do you want the engine in front???!
ha ha. Sony already dodged this problem with their petasite, which has been around for years.
Here's one of the first images done in 1926. Of course, this weather image didn't come from a satellite, but they've been doing this stuff for a long time!
And for some that actually is powered by something pulled off an R/C plane...
Actually, this is 300k baud - much faster than the //c can handle. If you wrote the code in assembly (running with a 1 MHz clock and 4 cycles per instruction), the tightest code ("LDA $C1FF; LDA $C1FF; etc.") wouldn't keep up -- and that doesn't even do anything with the data! Ah, nostalgia. I remember how the //c couldn't scroll the screen reliably at 2400 baud - it's amazing how far things have come, especially when I DMA stuff at a gigabit/sec to/from a RAID.
Thanks, Xerithane. My submission was in the queue for the longest time before it was rejected (I was keeping an eye on it because I realized I had misspelled benetton in the title). When it finally got rejected, I put in in my journal. It's still listed as rejected; Too bad people looking at my user profile will only see my journal entry and not the link to the larger discussion...
If the site is slashdotted, you can just download the full-resolution image [617.7 megabyte TIFF]
Way too complicated for 99% of the people!! It would be much better to invest a little more smarts in the infrastructure considering the complete lack of tech support that'll be available at the stores. Consider this scenario:
Everything is DHCP-friendly, so people don't have to change their network settings. However, all things go through a proxy computer - when it detects a new MAC address, it will return the sign-in webpage for any page you request. Until you enter the PIN from your receipt or a credit-card number, that will be the only page you'll get. (Or, as a bonus, maybe they'll let you browse a couple of other specific pages for free.)
Heck, you could also have the proxy spoof the email, too. When your computer thinks its logging into your corporate mailserver, it's really just logging into the macdonalds computer... and, surprise, no matter what name/password you use, you'll have some new mail... with a link to the mcdonalds sign-in page, of course!
Transportation- How do you think the food gets to your deli or bakery? In the old days, your choice of husband/wife was limited to about a half-day's walk from your home - if you lived in the boonies, then it's tough luck.
Neat story about the well & socializing, thanks!
Regular jobs are still in the majority; consultants are possible because their networking opportunites are greatly expanded with the use of a phone, a car, and sometimes a computer. My contracter friends swear by payroll services, again a "necessary" worksaver.
Like I said, some people live without many of these things; for others they find them far more beneficial than their costs.
Where I used to work, we sold gigabit fibre channel interfaces that ran at about the same speed. I was thinking of how to show off the technology at tradeshows -- we ultimately went with a 2048x1024ish video transfer at high frame rates, but I thought of having CD pass by (like on a carasol[sp]) at the rate their information was played at - 850X sounds much more impressive than gigabit. But, that wouldn't be eye-catching enough, so I thought of 3.5" floppies... but 85 floppies per second was likely to injure someone!!