IDSL is an interesting case. IDSL is broadband. ISDN isn't.
I'm afraid that's not true. The whole point of IDSL is that it uses existing baseband ISDN infrastructure out to the customer. You're just using the line as a dumb link to your ISP instead of switched, billed-per-minute access to the PSTN.
It's old-school transmission technology - just a bunch of baseband repeaters like at traditional T1 line.
"You must not maintain or permit multiple concurrent connections to the Internet Access."
Yes, god forbid I should have multiple concurrent conections to "the internet access". Does that mean I can't read the web while downloading something? Christ, every modern web browser uses at least two concurrent connections to download a page, usually four.
never recording Your password on Your computer
What a joke. So they've banned the feature in many web browsers which lets you save (auto-enter) login information for your low-security sites.
Because hopefully in the near future, MEMS "switches" will use tiny mirrors to create an ephemeral, purely optical, dedicated path across the world, through all the fibers directly to your home. Or maybe it's not dedicated, but a number of key hops along the way are "opened up" to massive capacity in response to your request. Once the path is set up (about 100ms), you have an uberfast link all thw way to the server. But it's expensive, so you have to use it and get off of it quick.
So you need a transfer protocol that can use that path as efficiently as possible. You download the movie in five seconds, it gets dumped into RAM, and then the link shrinks back to normal. The movie starts, while your PC spends the next couple minutes dumping the data from RAM to your disk.
Rambling on... I skimmed the article briefly - it explains what they accomplished in their test, but not how the protocol works. The key question is how "fair" is this protocol to regular TCP traffic? It's easy to make TCP go faster for bulk transfers by turning a few knobs at the sender's side (esp CWND backoff and growth parameters) or just opening multiple connections in parallel. But you're just stealing bandwidth from "regular" TCP connections when the link is saturated.
When dealing with "log fat pipes", even uncongested ones, the key is to grow the congestion window to the capacity of the link and then keep it there. TCP is very conservative and will back off too much when it hits the ceiling - then it takes a few RTTs to get back up to the link capacity. On a LAN it's no problem but at long distances it gets tricky.
Changing TCP so it uses the full capacity of an uncongested link, but also acts fairly in the event of congestion... that's hard. Would love to know how they did this.
The problem is that the fastest hard drives on the market today are Ultra320 SCSI, which have a throughput of 320MB per sec
That really depends on the application. A server with a lot of RAM that is cranking out a static web site will almost never hit the disk. LAN, CPU, disk - any one of these could be the bottleneck depending on what you're doing.
Also I must point out that a single Ultra320 disk isn't going to sustain throughput anywhere near the bus capacity of 320MB/s. I admit I haven't actually benchmarked a Ultra320 drive, but I expect you'd see something on the order of 40 or 50MB/s.
I didn't hear Saddam going on Al Jazeera (sp) on Sept 11th, condemning the horrific attack on civilians. He would love nothing more than to see us and Isreal exterminated. If that's not an Islamic militant then I don't know what is.
THIS is the sort of thing that is worthy of a patent.
Yes, I can see it now... "method and device for transmitting power by means of inductive coupling".
Here's a clue: the transformer been around for about as long as we've known about electricity.
ALso, the idea of using a transformer as the coupling for a charger base has been around for as long as I can remember. As a kid I had an electric toothbrush that did this. It's a great idea for something like a toothbrush, because it's harder to waterproof an electrical contact.
In high school we were doing some experiment in groups, spinning a cork on the end of string, with the string going through a tube and a weight on the bottom, to demonstrate centipetal force. I forget exactly what the exercise was, but it wasn't terribly interesting and everyone was just screwing around. I got bored and started checking out what was in the cabinets along the wall.
When the teacher wasn't looking, I pulled out a high voltage transformer and a few bits of heavy wire. I hadn't done this before, but I'd read about it in a.txt file I'd downloaded somewhere. Anyway I formed the wired and hooked them to the transformer and fired the thing up. It was great. In about 30 seconds half the class was crowded around this thing and asking me how it worked.
For the rest of the class, we ditched the centripetal force thing, and she had me at the white board explaining how a Jacob's Ladder works. I'll never forget it...
I'd love to be a science teacher some day. Sadly, teaching hardly pays a living wage in California, so instead I sit at a desk writing code.
As long as we're burning fossil fuels to generate power, all an electric car does is move the pollution somewhere else. Just think about it:
Gas car: Chemical energy -> kinetic energy
Electric car: Chemical energy -> kinetic -> electrical -> long distance transmission (power lines) -> chemical (batteries) -> electrical -> kinetic
In the end, you get sucky performance for a couple times the energy cost. The idea of an electric car is utterly absurd, and I can't understand why it happened at all.
Maybe after get serious about cheap, clean nuclear power, and we make some major breakthroughs in batteries, the electric car can happen.
Hmmm... I remember a story about the tards at Salon.com who had a long-term lease on some $20+/ft office space, but there aren't too many companies paying that kind of $$ here. Office space in the south bay is down from $6.00/ft to about $2.00/ft for a very nice full-service space. Larger offices (10K+ sq ft) are well under a buck. Industrial/commercial space w/o cleaning and utilities can be had for $0.60.
As a young guy trying to make it here, there is definitely a sense a schadenfreude in seeing the super-rich commercial property owners getting the tables turned after gouging every business in the valley. But I'm still burning $1500/mo here for a small 2br appartment. A 30% dip in housing prices would go a long way towards putting things back into balance, but I'm afraid it's not going to happen.
Tpropety owners here are just absurdly well-off. Generally, the guys who own all the commercial space can easily afford to just leave it vacant for another five years if they have to.
int a, b;
typedef int t, u;
void f1() { a * b; }
void f2() { t * u; }
void f3() { t * b; }
void f4() { int t; t * b; }
void f5(t u, unsigned t) {
switch ( t ) {
case 0: if ( u )
default: return;
}
}
Also why is such code used in the kernel? I know there are practical reasons for things like "do {} while (0)" but the code above just looks deliberately obfuscated.
I'm wondering if they did this using TCP, and if so, how the test was set up.
Eg is this the peak rate that it was able to sustain for a one minute period once the transfer was already running, or did it take one minute from start to finish. It's an important distinction with TCP because slow start needs several round trips in order to open the window large enough to get max througput over such a high speed, long distance link.
Also how on earth did they handle packet loss? Getting the max throughput out of a high-latency link with just a single TCP connection is not easy.
I bought their Optra C710 color laser about a year. I thought I was getting a deal at $1200, but it is the worst printer I have ever owned.
It came with all toner cartridges only 25% filled. This was not mentioned anywhere on the box or on the web site where I ordered.
The printer has actually functioned maybe half of the time that we've owned it. Two on-site service calls later, and we're still having problems:
Why does it say paper jam when there is no paper jam?
"Coating roll life warning" and "transfer belt warning" come up all the time, even right after fresh ones are installed
Duplexing option jams on every 100th sheet
Print often seems to stick to the transfer belt and gets "ghosted" onto subsequent pages
The printer just disappears from the network at least once per day and needs a hard reboot
In contrast, our HP laserjet has NEVER missed a beat. Look I know this is not a representative sample or anything, but there are clearly DESIGN flaws with this printer and it should not be on the market. Even after multiple service calls it does not work.
I fear that there is NO profitable model ! Any model that requires payment is doomed because it can not compete with Kazaa-like stuff.
I disagree. I'll pay for quality tracks and accurate ID3 tags. It's worth it if I can get good filez directly from a fast server instead of scrounging the p2p scene.
Slashdot Mirror
Ah diet pepsi. All the caffeine, but without that nasty sugar.
IDSL is an interesting case. IDSL is broadband. ISDN isn't.
I'm afraid that's not true. The whole point of IDSL is that it uses existing baseband ISDN infrastructure out to the customer. You're just using the line as a dumb link to your ISP instead of switched, billed-per-minute access to the PSTN.
It's old-school transmission technology - just a bunch of baseband repeaters like at traditional T1 line.
"You must not maintain or permit multiple concurrent connections to the Internet Access."
Yes, god forbid I should have multiple concurrent conections to "the internet access". Does that mean I can't read the web while downloading something? Christ, every modern web browser uses at least two concurrent connections to download a page, usually four.
never recording Your password on Your computer
What a joke. So they've banned the feature in many web browsers which lets you save (auto-enter) login information for your low-security sites.
How gay.
Only 2 symbols are required to express any algorythm (1 and 0)
No. One also needs a processor/compiler/interpreter. A string of bits is meaningless per se.
Browsing an iPhoto library with 2000+ files strikingly slow, and surprisingly fast considering the math that is going into it.
That's just because iPhoto is retarded. I mean, how hard is it to do that once, and then cache the resulting thumbnail for later?
Why would you need to download it in 5 seconds?
Because hopefully in the near future, MEMS "switches" will use tiny mirrors to create an ephemeral, purely optical, dedicated path across the world, through all the fibers directly to your home. Or maybe it's not dedicated, but a number of key hops along the way are "opened up" to massive capacity in response to your request. Once the path is set up (about 100ms), you have an uberfast link all thw way to the server. But it's expensive, so you have to use it and get off of it quick.
So you need a transfer protocol that can use that path as efficiently as possible. You download the movie in five seconds, it gets dumped into RAM, and then the link shrinks back to normal. The movie starts, while your PC spends the next couple minutes dumping the data from RAM to your disk.
Rambling on... I skimmed the article briefly - it explains what they accomplished in their test, but not how the protocol works. The key question is how "fair" is this protocol to regular TCP traffic? It's easy to make TCP go faster for bulk transfers by turning a few knobs at the sender's side (esp CWND backoff and growth parameters) or just opening multiple connections in parallel. But you're just stealing bandwidth from "regular" TCP connections when the link is saturated.
When dealing with "log fat pipes", even uncongested ones, the key is to grow the congestion window to the capacity of the link and then keep it there. TCP is very conservative and will back off too much when it hits the ceiling - then it takes a few RTTs to get back up to the link capacity. On a LAN it's no problem but at long distances it gets tricky.
Changing TCP so it uses the full capacity of an uncongested link, but also acts fairly in the event of congestion... that's hard. Would love to know how they did this.
The problem is that the fastest hard drives on the market today are Ultra320 SCSI, which have a throughput of 320MB per sec
That really depends on the application. A server with a lot of RAM that is cranking out a static web site will almost never hit the disk. LAN, CPU, disk - any one of these could be the bottleneck depending on what you're doing.
Also I must point out that a single Ultra320 disk isn't going to sustain throughput anywhere near the bus capacity of 320MB/s. I admit I haven't actually benchmarked a Ultra320 drive, but I expect you'd see something on the order of 40 or 50MB/s.
BTW, Saddam is not an Islamic militant.
I didn't hear Saddam going on Al Jazeera (sp) on Sept 11th, condemning the horrific attack on civilians. He would love nothing more than to see us and Isreal exterminated. If that's not an Islamic militant then I don't know what is.
What I'm trying to figure out is this:
if we go in through Turkey and take Iraq from behind, would Greece help?
THIS is the sort of thing that is worthy of a patent.
Yes, I can see it now... "method and device for transmitting power by means of inductive coupling".
Here's a clue: the transformer been around for about as long as we've known about electricity.
ALso, the idea of using a transformer as the coupling for a charger base has been around for as long as I can remember. As a kid I had an electric toothbrush that did this. It's a great idea for something like a toothbrush, because it's harder to waterproof an electrical contact.
In high school we were doing some experiment in groups, spinning a cork on the end of string, with the string going through a tube and a weight on the bottom, to demonstrate centipetal force. I forget exactly what the exercise was, but it wasn't terribly interesting and everyone was just screwing around. I got bored and started checking out what was in the cabinets along the wall.
.txt file I'd downloaded somewhere. Anyway I formed the wired and hooked them to the transformer and fired the thing up. It was great. In about 30 seconds half the class was crowded around this thing and asking me how it worked.
When the teacher wasn't looking, I pulled out a high voltage transformer and a few bits of heavy wire. I hadn't done this before, but I'd read about it in a
For the rest of the class, we ditched the centripetal force thing, and she had me at the white board explaining how a Jacob's Ladder works. I'll never forget it...
I'd love to be a science teacher some day. Sadly, teaching hardly pays a living wage in California, so instead I sit at a desk writing code.
I wonder if Open Source could contribute to an economic comeback in any way.
Gee I wonder... what powered that whole "Internet" craze? Sure as hell wasn't Windows 3.1.
If OSS had a ticker symbol, I'd buy in a heartbeat.
copyrightable[1]
[1] Is that really a word?
Perfectly cromulant in my book.
Nuclear
As long as we're burning fossil fuels to generate power, all an electric car does is move the pollution somewhere else. Just think about it:
Gas car: Chemical energy -> kinetic energy
Electric car: Chemical energy -> kinetic -> electrical -> long distance transmission (power lines) -> chemical (batteries) -> electrical -> kinetic
In the end, you get sucky performance for a couple times the energy cost. The idea of an electric car is utterly absurd, and I can't understand why it happened at all.
Maybe after get serious about cheap, clean nuclear power, and we make some major breakthroughs in batteries, the electric car can happen.
which guarantees a +5 moderation whenever you say:
feel free to moderate me down into oblivion
Seriously, I've never seen this except on +5 comments. Feel free to moderate me down into oblivion.
Yeah, but it'll take passwords like 123!@#qwe!@#
Hint: look at your keyboard.
Hmmm... I remember a story about the tards at Salon.com who had a long-term lease on some $20+/ft office space, but there aren't too many companies paying that kind of $$ here. Office space in the south bay is down from $6.00/ft to about $2.00/ft for a very nice full-service space. Larger offices (10K+ sq ft) are well under a buck. Industrial/commercial space w/o cleaning and utilities can be had for $0.60.
As a young guy trying to make it here, there is definitely a sense a schadenfreude in seeing the super-rich commercial property owners getting the tables turned after gouging every business in the valley. But I'm still burning $1500/mo here for a small 2br appartment. A 30% dip in housing prices would go a long way towards putting things back into balance, but I'm afraid it's not going to happen.
Tpropety owners here are just absurdly well-off. Generally, the guys who own all the commercial space can easily afford to just leave it vacant for another five years if they have to.
15 years ago I moved out of the Valley (Saratoga) because it (SV) was turning into a overcrowded cesspool.
I'll bet you just couldn't make the payments on your $24M two-bedroom home.
I'm wondering if they did this using TCP, and if so, how the test was set up.
Eg is this the peak rate that it was able to sustain for a one minute period once the transfer was already running, or did it take one minute from start to finish. It's an important distinction with TCP because slow start needs several round trips in order to open the window large enough to get max througput over such a high speed, long distance link.
Also how on earth did they handle packet loss? Getting the max throughput out of a high-latency link with just a single TCP connection is not easy.
go RTFA.
If you're having trouble with acronyms, go look them up at everything2
TTYL.
It came with all toner cartridges only 25% filled. This was not mentioned anywhere on the box or on the web site where I ordered.
The printer has actually functioned maybe half of the time that we've owned it. Two on-site service calls later, and we're still having problems:
In contrast, our HP laserjet has NEVER missed a beat. Look I know this is not a representative sample or anything, but there are clearly DESIGN flaws with this printer and it should not be on the market. Even after multiple service calls it does not work.
I fear that there is NO profitable model ! Any model that requires payment is doomed because it can not compete with Kazaa-like stuff.
I disagree. I'll pay for quality tracks and accurate ID3 tags. It's worth it if I can get good filez directly from a fast server instead of scrounging the p2p scene.
Er... meant to say:
"the USA is starting this war?"
You don't like the USA military starting a war with Iraq.
Iraq's starting this war? Have you been in a fucking coma since 09/10/01?
War is a fact of life. If you're not willing to fight, someone else is, and YOU WILL DIE.