The price structure represents more than just the predicted amount of traffic going over the wire. It also has to do with the value of the line, as well as some of the politics that go into regulating a monopoly.
Quite simply, businesses get alot more monetary benefit out of having a phone line than do home users. Most businesses wouldn't be in business if they didn't have a phone line nowadays, since they would have a difficult time running the business. So, having a phone line is very valuable to the business. Businesses also have deeper pockets, so it seems justified that a business line cost more than a home line.
Residential lines are more convenient than anything. People won't be "going out of business" if they don't have phones. So, from that standpoint, phone lines don't carry the monetary benefit for residential customers that they do for businesses. So generally, residential service is less valuable than business service, and so it makes sense to charge less. Also, telephones provide basic access to emergency services as well, which is where politics come in: Since Telcos tend to operate as monopolies or near-monopolies in most regions, and since telephone access is needed for services such as 911, telcos need to offer lower-than-usual rates for the low end of the cost spectrum.
There's a third argument which sews up the business vs. residential pricing policy rather nicely as well: The major reason that business phones are so valuable is so that customers can call the business. Most businesses serve the public-at-large, and so the phones are only truly valuable if phones are widely deployed amongst the public. Yet another reason for the cost disparity.
Regional Bell Operating Company, if memory serves. It's the official name for the Baby Bells, if I'm recalling correctly. If not, then could someone please smack me with a clue stick?
We're very quickly entering an era where computers with Internet access are no longer a luxury, but are a requirement to access certain goods and services. Consider, for example, Arizona's DMV (I think it was Arizona) that is offering services over the web. I'm sure it would love to offer services over the web exclusively if it could, with web-enabled kiosks in place of the old public offices for people who can't get a computer of their own. While you're not required to own a computer with a web browser in such a world, it would be very much in your own interest to have one. (Much like a car -- there's no explicit requirement to own a car, but most people feel they need one because life without one is pretty painful, at least in most places in the US.)
Lets not forget Voice over IP and other such technologies either: Your only line might be a data line.
On a separate angle, consider that traditional ISPs are already starting to feel the heat from cable modem and DSL services -- neither of which affects the capacity of the POTS network and therefore would likely be exempt from the proposed "tax". Instituting a tax such as this would only make life worse for these traditional ISPs, possibly slowly forcing them out of business. Where does the business go? At least some of these customers will upgrade to DSL or cable-modem when their current ISP goes away.
Who's providing the DSL and the cable-modems? Phone and cable companies -- and these days, the line between the two is getting blurrier and blurrier. OpTel, my cable company (and ISP, since I have a cable modem), would love to be my phone company as well. (They offer both phone and cable services in my apartment complex.) Also, doesn't AT&T have a stake in TCI? It would be interesting to see the list of companies that are lobbying for this tax, and what they stand to gain as traditional ISPs fall into obscurity/insolvency.
Yes, I read the rules, and I realize that the order of submission doesn't matter a whole lot. But, there's a nice bit of ego involved in knowing I had the right answer before someone posted the algorithm to/. for the whole world to see.
The ISBN is important, yes. In calculation, no. (I'm pretty sure I just solved the cypher. The message was twisted enough that it just might be right.)
I'm sorry you didn't like TRON. It was one of my favorite childhood computer movies. It does a good job of keeping the computers abstracted through most of the "real world" scenes, and the "in-the-computer" portion is truly more a fantasy adventure anyway.
Granted, a full 3-D video game that isn't vector-based in 1982 is beyond unlikely, but still, the movie was pretty reasonable overall. Also, what other movie features someone cracking from an Apple///? (You read that right, an Apple///, as in 3. You get to see it when Flynn is "talking" to CLU.)
BTW, for you TRON fans out there... There's a neat article in The Onion: Our Dumb Century entitled: "Reagan to Meet with Master Control Program". Yes, that Master Control Program. It even mentions Ed Dillinger and ENCOM. (On the previous page, a headline reads "Secret Pac-Man Patterns Fall into Russian Hands." Buy the book, darnit!)
I know that the intro to Office Space shows a brief clip of east-bound traffic on I-635 (LBJ Freeway) near the Dallas North Tollway. Also, Peter's neighbor Lawrence mentions his current construction job is at the McDonalds in Los Colinas (which is between Dallas and DFW Airport).
Now I know Mike Judge sets just about everything in Texas, and usually near Dallas. (For instance, King of the Hill is based in "Arlen," Texas, which does not exist -- but compare to Garland which is a Dallas suburb.) Does anyone here know where Office Space was filmed?
It's worse than that. The pathnames have UNIX-style forward slashes much of the time (but if I recall correctly, not all of the time). And when Peter Gibbons is playing Tetris, it's the X-Windows Tetris (I'm pretty sure complete w/ Motif widgets). To top it off, the C:> prompt is missing the slash entirely, reminiscent of the DOS 2.x-era default PROMPT.
For instance, have you ever seen a Middle Eastern man in a movie who was not a terrorist?
Sure! Have you seen Office Space yet? Samir (I'm pretty sure that was his name) wasn't a terrorist... (although he stood a chance of going to Federal Pound-Me-In-The-Ass Prison...)
Microsoft's flight sim was acquired from SubLogic, wasn't it?
Also, on the "bigger is better" theme: Yes, Microsoft and every other US retailer seems to benefit from that psychology. Pontiac sold me a nice, supercharged Grand Prix with nifty blinkenlights to tell me how much boost I'm getting. So what? I don't see how that's relevant to Linux vs. Microsoft here. There are Linux distributions that are every bit as bloated as Microsoft's Windows (for instance, RedHat, which I run at home).
(That reminds me, I really must give Debian a whirl. That is, as long as I remember my waders before I start traipsing through the thick field GPL religious dogma that surrounds it. I respect the GPL and I'm even distributing my own code under it, but I think it's possible to take an ideology too far. I view Free Software as a means to encourage Good Engineering, not as Theology and Religion.)
I like coding by NIN, although it can have some interesting side effects.
I was writing one program while listening to broken and fixed. In part of my code, I had routines for allocating and deallocating objects of a particular type. In the allocator, comments to the effect of "I want you to make me, I want you to take me" snuck in, with the corresponding "I want you to break me, I want you to throw me away" in the deallocator....:-)
NIN is good (particularly The Fragile, broken and fixed, and the downward spiral), since there is a good mix of hard, heavy pounding music, and softer, calmer music.
The speed of light in a vaccuum is constant. Electric current running down a wire is in anything but a vacuum, however. The rate at which a electrical signal propogates in a given wire has to do with the series inductance of the wire and the parallel capacitance of the wire, per unit of length. (Parallel to the shielding, that is. Unshielded wires are often treated as having shielding at "infinity" that is tied to ground.) The ratio of these quantities forms the "impedance" of the wire, which is usually specified in Ohms, and is invariant with respect to length. That's why, regardless of its length, RG-6 Coax (cable TV wire) is 75ohm.
(Note that I'm talking about ideal wires here. The series resistance of real wires also mucks up the propogation rate somewhat. Not only that, but also it delays up signals at different frequencies differently, leading to something known as "skew", if I recall correctly. Whee. Transmission line theory was never my strong point, but I'll never forget the basic lessons it taught.)
Ack... HDLC bitstuffing... it's a nightmare. Also, it makes your channel's bandwidth vary by up to 20%. Bleh...
(For the uninitiated, HDLC framing requires a 0 be inserted in the bit-stream after five consecutive 1s, to allow for out-of-band messages and timing recovery. (Six consecutive 1s have their own special, out-of-band meaning.) In the worst case, a bit stream of all 1s expands by a ratio of 6/5.)
Indeed, as far back as 2400bps modems, the difference between bits per second differed from baud. If I recall correctly, the QPSK scheme 2400bps uses sends two bits per transition, and so is 1200 baud, 2400 bps.
One interesting thing to note is the fact that most modulation schemes for modern modem hardware mostly eliminate the stop-bit/start-bit/parity overhead that RS232 has. This started with the 2400bps error correcting modems Way Back When. On these modems, ZModem download rates as high as ~275 bytes/second were not uncommon, since the modems disassociated the RS232 signal from the audio signalling. In contrast, the old-fashioned Bell 110 "0 - 600 baud" modems were little more than a voltage-controlled oscillator hooked to the Tx line and PLLs hooked to the Rx line and Carrier Detect line.
Back to the original topic: Transports (such as Ethernet, T1s, etc.) are specified in Bits per Second largely because they are a bit-oriented medium, and they generally have little protocol overhead associated with them. (Although T1s do implement "bit robbing" for their control channel, thus reducing their advertised capacity to 1.536Mbps rather than 1.544Mbps.) Higher-order operations such as file transfers are reported in Bytes per Second, since (1) files are usually collections of bytes, and (2) the protocol stack provides a byte-oriented interface to the file transfer program. Somewhere in the stack (usually very, very near the bottom, at the Physical layer), the bytes become bits (and vice versa), but by that time you've inherited all of the protocol overhead of whatever protocols you're running (anything from ZModem to TCP/IP to SNA to whatever) and so bits vs. bytes starts looking alot like apples vs. oranges.
It appears that idSoftware has stressed portability since about the Doom era. I recall John Carmack stating that Doom ported to Jaguar in a weekend (although it wasn't terribly efficient at first). I've even ported Doom to one of TI's DSPs in my copious free time. How's that for portable?
I personally guess the answer is one of the following: He's already happy with his income level, and he realizes that Doom etc. wouldn't be nearly as popular (at the retailers) without the level of exposure the piracy has gotten him.
To be sure, Wolfenstein was a sort of 2-D raytracing (for wall discovery and texture selection). A separate ray was shot on the 2-D map grid for every column of the display. Where the ray hit determined the texture column and height to draw in that column of the display.
Personally, I feel that full 3-D Povray-style texture mapping will probably not be feasible until special parallel raytrace processing units. By that time, general processing should be fast enough to do something better anyway (such as radiosity-based rendering).
The May 1999 issue of Car & Driver had an article entitled Doctor Yes and the Mad Ferarris. This article covered both your Testarossa and your F50, as well as your attempts to break the 200MPH mark.
First, let me say, I've driven by (I live in the area) and those are mighty beautiful machines. Have you managed to take either of these gorgeous rockets past 200MPH since the article?
I emailed John Markoff. Apparently, the copy desk at NYTimes fubared this one -- it's not his fault. (I wrote him a nice friendly email, and he responded fairly quickly, actually.)
I've got a 300MHz Ultra 2 on my desk at work, and a Pentium II 300MHz at home. The Ultra is alot faster at batch floating point jobs, but the Pentium II running Linux feels alot snappier than the Ultra 2 running Solaris 2.5.1. In general, the Ultra 2 is faster on big batch jobs though. X Windows performance sucks though, since my workstation seems to have a minimally accelerated frame-buffer.
Slashdot Mirror
There was a Slashdot Article on September 4th.
The world lost a truly great networking mind when it lost W. Richard Stevens. May he rest in peace.
--Joe--
The price structure represents more than just the predicted amount of traffic going over the wire. It also has to do with the value of the line, as well as some of the politics that go into regulating a monopoly.
Quite simply, businesses get alot more monetary benefit out of having a phone line than do home users. Most businesses wouldn't be in business if they didn't have a phone line nowadays, since they would have a difficult time running the business. So, having a phone line is very valuable to the business. Businesses also have deeper pockets, so it seems justified that a business line cost more than a home line.
Residential lines are more convenient than anything. People won't be "going out of business" if they don't have phones. So, from that standpoint, phone lines don't carry the monetary benefit for residential customers that they do for businesses. So generally, residential service is less valuable than business service, and so it makes sense to charge less. Also, telephones provide basic access to emergency services as well, which is where politics come in: Since Telcos tend to operate as monopolies or near-monopolies in most regions, and since telephone access is needed for services such as 911, telcos need to offer lower-than-usual rates for the low end of the cost spectrum.
There's a third argument which sews up the business vs. residential pricing policy rather nicely as well: The major reason that business phones are so valuable is so that customers can call the business. Most businesses serve the public-at-large, and so the phones are only truly valuable if phones are widely deployed amongst the public. Yet another reason for the cost disparity.
--Joe--
Regional Bell Operating Company, if memory serves. It's the official name for the Baby Bells, if I'm recalling correctly. If not, then could someone please smack me with a clue stick?
--Joe--
We're very quickly entering an era where computers with Internet access are no longer a luxury, but are a requirement to access certain goods and services. Consider, for example, Arizona's DMV (I think it was Arizona) that is offering services over the web. I'm sure it would love to offer services over the web exclusively if it could, with web-enabled kiosks in place of the old public offices for people who can't get a computer of their own. While you're not required to own a computer with a web browser in such a world, it would be very much in your own interest to have one. (Much like a car -- there's no explicit requirement to own a car, but most people feel they need one because life without one is pretty painful, at least in most places in the US.)
Lets not forget Voice over IP and other such technologies either: Your only line might be a data line.
On a separate angle, consider that traditional ISPs are already starting to feel the heat from cable modem and DSL services -- neither of which affects the capacity of the POTS network and therefore would likely be exempt from the proposed "tax". Instituting a tax such as this would only make life worse for these traditional ISPs, possibly slowly forcing them out of business. Where does the business go? At least some of these customers will upgrade to DSL or cable-modem when their current ISP goes away.
Who's providing the DSL and the cable-modems? Phone and cable companies -- and these days, the line between the two is getting blurrier and blurrier. OpTel, my cable company (and ISP, since I have a cable modem), would love to be my phone company as well. (They offer both phone and cable services in my apartment complex.) Also, doesn't AT&T have a stake in TCI? It would be interesting to see the list of companies that are lobbying for this tax, and what they stand to gain as traditional ISPs fall into obscurity/insolvency.
--Joe--
Yes, I read the rules, and I realize that the order of submission doesn't matter a whole lot. But, there's a nice bit of ego involved in knowing I had the right answer before someone posted the algorithm to /. for the whole world to see.
--Joe--
I wonder if my solution (all of which was grammatically correct) made it in before the /. effect hit. *sigh*
--Joe--
The ISBN is important, yes. In calculation, no. (I'm pretty sure I just solved the cypher. The message was twisted enough that it just might be right.)
--Joe--
I'm sorry you didn't like TRON. It was one of my favorite childhood computer movies. It does a good job of keeping the computers abstracted through most of the "real world" scenes, and the "in-the-computer" portion is truly more a fantasy adventure anyway.
Granted, a full 3-D video game that isn't vector-based in 1982 is beyond unlikely, but still, the movie was pretty reasonable overall. Also, what other movie features someone cracking from an Apple /// ? (You read that right, an Apple ///, as in 3. You get to see it when Flynn is "talking" to CLU.)
BTW, for you TRON fans out there... There's a neat article in The Onion: Our Dumb Century entitled: "Reagan to Meet with Master Control Program". Yes, that Master Control Program. It even mentions Ed Dillinger and ENCOM. (On the previous page, a headline reads "Secret Pac-Man Patterns Fall into Russian Hands." Buy the book, darnit!)
--Joe--
I know that the intro to Office Space shows a brief clip of east-bound traffic on I-635 (LBJ Freeway) near the Dallas North Tollway. Also, Peter's neighbor Lawrence mentions his current construction job is at the McDonalds in Los Colinas (which is between Dallas and DFW Airport).
Now I know Mike Judge sets just about everything in Texas, and usually near Dallas. (For instance, King of the Hill is based in "Arlen," Texas, which does not exist -- but compare to Garland which is a Dallas suburb.) Does anyone here know where Office Space was filmed?
--Joe--
It's worse than that. The pathnames have UNIX-style forward slashes much of the time (but if I recall correctly, not all of the time). And when Peter Gibbons is playing Tetris, it's the X-Windows Tetris (I'm pretty sure complete w/ Motif widgets). To top it off, the C:> prompt is missing the slash entirely, reminiscent of the DOS 2.x-era default PROMPT.
--Joe--
Go rent the movie, re-read his comment, and realize he was referring to a particularly hilarious running gag from it.
--Joe--
Sure! Have you seen Office Space yet? Samir (I'm pretty sure that was his name) wasn't a terrorist... (although he stood a chance of going to Federal Pound-Me-In-The-Ass Prison...)
--Joe--
Microsoft's flight sim was acquired from SubLogic, wasn't it?
Also, on the "bigger is better" theme: Yes, Microsoft and every other US retailer seems to benefit from that psychology. Pontiac sold me a nice, supercharged Grand Prix with nifty blinkenlights to tell me how much boost I'm getting. So what? I don't see how that's relevant to Linux vs. Microsoft here. There are Linux distributions that are every bit as bloated as Microsoft's Windows (for instance, RedHat, which I run at home).
(That reminds me, I really must give Debian a whirl. That is, as long as I remember my waders before I start traipsing through the thick field GPL religious dogma that surrounds it. I respect the GPL and I'm even distributing my own code under it, but I think it's possible to take an ideology too far. I view Free Software as a means to encourage Good Engineering, not as Theology and Religion.)
--Joe--
Rather than Kool-Aid, they would've used cyanide-laced Penguin Mints and Jolt.
--Joe--
I like coding by NIN, although it can have some interesting side effects.
I was writing one program while listening to broken and fixed. In part of my code, I had routines for allocating and deallocating objects of a particular type. In the allocator, comments to the effect of "I want you to make me, I want you to take me" snuck in, with the corresponding "I want you to break me, I want you to throw me away" in the deallocator.... :-)
NIN is good (particularly The Fragile, broken and fixed, and the downward spiral), since there is a good mix of hard, heavy pounding music, and softer, calmer music.
--Joe--
The speed of light in a vaccuum is constant. Electric current running down a wire is in anything but a vacuum, however. The rate at which a electrical signal propogates in a given wire has to do with the series inductance of the wire and the parallel capacitance of the wire, per unit of length. (Parallel to the shielding, that is. Unshielded wires are often treated as having shielding at "infinity" that is tied to ground.) The ratio of these quantities forms the "impedance" of the wire, which is usually specified in Ohms, and is invariant with respect to length. That's why, regardless of its length, RG-6 Coax (cable TV wire) is 75ohm.
(Note that I'm talking about ideal wires here. The series resistance of real wires also mucks up the propogation rate somewhat. Not only that, but also it delays up signals at different frequencies differently, leading to something known as "skew", if I recall correctly. Whee. Transmission line theory was never my strong point, but I'll never forget the basic lessons it taught.)
--Joe--
Ack... HDLC bitstuffing... it's a nightmare. Also, it makes your channel's bandwidth vary by up to 20%. Bleh...
(For the uninitiated, HDLC framing requires a 0 be inserted in the bit-stream after five consecutive 1s, to allow for out-of-band messages and timing recovery. (Six consecutive 1s have their own special, out-of-band meaning.) In the worst case, a bit stream of all 1s expands by a ratio of 6/5.)
--Joe--
Indeed, as far back as 2400bps modems, the difference between bits per second differed from baud. If I recall correctly, the QPSK scheme 2400bps uses sends two bits per transition, and so is 1200 baud, 2400 bps.
One interesting thing to note is the fact that most modulation schemes for modern modem hardware mostly eliminate the stop-bit/start-bit/parity overhead that RS232 has. This started with the 2400bps error correcting modems Way Back When. On these modems, ZModem download rates as high as ~275 bytes/second were not uncommon, since the modems disassociated the RS232 signal from the audio signalling. In contrast, the old-fashioned Bell 110 "0 - 600 baud" modems were little more than a voltage-controlled oscillator hooked to the Tx line and PLLs hooked to the Rx line and Carrier Detect line.
Back to the original topic: Transports (such as Ethernet, T1s, etc.) are specified in Bits per Second largely because they are a bit-oriented medium, and they generally have little protocol overhead associated with them. (Although T1s do implement "bit robbing" for their control channel, thus reducing their advertised capacity to 1.536Mbps rather than 1.544Mbps.) Higher-order operations such as file transfers are reported in Bytes per Second, since (1) files are usually collections of bytes, and (2) the protocol stack provides a byte-oriented interface to the file transfer program. Somewhere in the stack (usually very, very near the bottom, at the Physical layer), the bytes become bits (and vice versa), but by that time you've inherited all of the protocol overhead of whatever protocols you're running (anything from ZModem to TCP/IP to SNA to whatever) and so bits vs. bytes starts looking alot like apples vs. oranges.
--Joe--
It appears that idSoftware has stressed portability since about the Doom era. I recall John Carmack stating that Doom ported to Jaguar in a weekend (although it wasn't terribly efficient at first). I've even ported Doom to one of TI's DSPs in my copious free time. How's that for portable?
--Joe--
I personally guess the answer is one of the following: He's already happy with his income level, and he realizes that Doom etc. wouldn't be nearly as popular (at the retailers) without the level of exposure the piracy has gotten him.
--Joe--
To be sure, Wolfenstein was a sort of 2-D raytracing (for wall discovery and texture selection). A separate ray was shot on the 2-D map grid for every column of the display. Where the ray hit determined the texture column and height to draw in that column of the display.
Personally, I feel that full 3-D Povray-style texture mapping will probably not be feasible until special parallel raytrace processing units. By that time, general processing should be fast enough to do something better anyway (such as radiosity-based rendering).
--Joe--
The May 1999 issue of Car & Driver had an article entitled Doctor Yes and the Mad Ferarris. This article covered both your Testarossa and your F50, as well as your attempts to break the 200MPH mark.
First, let me say, I've driven by (I live in the area) and those are mighty beautiful machines. Have you managed to take either of these gorgeous rockets past 200MPH since the article?
Also, could you autograph my copy of C&D?
--Joe--
I emailed John Markoff. Apparently, the copy desk at NYTimes fubared this one -- it's not his fault. (I wrote him a nice friendly email, and he responded fairly quickly, actually.)
--Joe--
I've got a 300MHz Ultra 2 on my desk at work, and a Pentium II 300MHz at home. The Ultra is alot faster at batch floating point jobs, but the Pentium II running Linux feels alot snappier than the Ultra 2 running Solaris 2.5.1. In general, the Ultra 2 is faster on big batch jobs though. X Windows performance sucks though, since my workstation seems to have a minimally accelerated frame-buffer.
--Joe--
Ack. Preview mode bit me. It turned my instances of <BR> into <BR> in my edit buffer and I didn't catch them all. Ack, ack, ack!
I meant to say "you can put a <BR> tag in there too".
--Joe--