You realize that you and I are part of the media right now, don't you?
In the example I cited, the FBI didn't make its opposition research available to the voters. Hundreds of background files somehow, mysteriously, ended up in an office in the White House.
It's not about whether the FBI monitors little old me. (Privacy advocates, hold on for a minute -- I care about little old me, too, but Gravis Zero doesn't).
Sun paid about $10 million to SCO and received warrants to buy 210,000 shares of SCOX as part of the deal.
I do agree with you that Sun has also done some awfully nice things for the open-source world. Sun is a friend when it's in Sun's interest (buying StarOffice and releasing the source under GPL) and a foe when that is in Sun's interest (funding SCO).
Pilot Richard Smith blacked out at the helm. Captain Michael Gansas was absent -- his duty station was in the forward wheelhouse to backup Smith, but he wasn't there. As a result of this double human failure, 11 people died.
Humans do fail, and so do computers. I don't think Smith's weakness was excusable but I have sympathy for him. Humans do fail, excusably or inexcusably. That's why the DOT pays a second human pilot to be in the wheelhouse during docking, but Gansas wasn't even there.
I would love for my wristwatch to pop up an alert telling me the optimal time to leave my office.
However, I think the MTA would take a dim view of:
-- Lots more people waiting to the literal last minute to catch a train, and then delaying 30 seconds more, and then running down the stairs.
-- Increased train dwell times from people mis-optimizing their platform arrival experience and forcing the doors open longer.
-- People noticing when trains are significantly late, or even just insignificantly late, and bitching about it, making life unpleasant for the MTA overlords.
Unfortunately, it's human nature for people to be dumbasses, and it's also human nature for large organizations to protect themselves against both dumbasses and smart complainers by hoarding information.
Also I think you've overestimated the positive benefits. If you want to be sure of catching your train, arrive 2 minutes early. I don't know about your job, but with my job, I actually do a lot of thinking away from my desk, and I use those several minutes per day for offline thinking.
A last thought: maybe the customers could do this P2P style with little or no co-operation from MTA. Imagine lots of people carrying PDA's that know where they are in the subway system and can upload real-time data streams to a commercial site. (That would be a badass PDA, and the app would have to run mostly in the background). The commercial site provides the information to your PDA from current riders. You either have to pay money, or watch ads, or leave your own PDA operational during your trip, or some combination of those things.
I've lost $0 in credit card theft, but dealing with the attempts has cost me quite a bit more in time: more than 100 times as much, in time and lost income, than the $80 cash loss.
Cash means more privacy, fewer entries to reconcile, and less paper to shred.
First, it was kind of silly to name the program the same as my user name, but I never found a better name for it than "my trace-and-replay debugger".
My original plan was to write this for Solaris and sell it, hence the insistence on tracing and replaying without modifying the target program or the operating system, and that's why the replay controller messes with gdb's mind, so that it can work with a stock gdb rather than needing gdb extensions.
I developed a Linux version first because it was so cheap and simple to throw Slackware on my computer. And, well, it turns out that I'll probably never need a Solaris version, because Linux sure has become big enough and rewarding enough for me.
Versions 0.1, 0.2, and 0.3 took 15 months of full-time work, living on my savings and writing code on a little Linux box.
After version 0.3, several bad things happened:
Technical butterfly-chasing: the tracer needs to know about all possible ioctl calls that the target program makes, and Linux was adding and changing ioctls faster than I could check the patch diffs to update them. (That's how I came to write the kernel change summaries for a while). The obvious solution is to trace about 20 common ioctl's and throw all the rest in a big worst-case box that says "this ioctl might touch all of memory". One of the problems of working alone: nobody else around to notice the obvious solution.
Moving up the management chain: anybody who runs a one-person operation knows about this. It's one thing to write a proof-of-concept program. It's another thing to push it out, start a community, market it, manage all the communication with users and co-developers. I failed at that.
Fade-out: after the proof-of-concept worked, I noticed I'd spent 15 months full time, and I did make the milestone of seeing test programs run. I lost some interest and went and did something else.
Version 0.3 still has one good use: to help defend against anybody else that files a patent for technology like this. I released in November 1995.
Some responses:
Zogger and Animats, that's exactly the use case, the user in the field runs the tracer, then mails a big log back to the developer. This gets very useful when the user has unique resources that the developer doesn't, for a program like a network server. I don't know much about dtrace, but I think dtrace is just more comprehensive kernel reporting information, not fundamentally "video-taping the user process".
Mebane, I think the answer is: in 1995, I sucked at explaining things to people. Specifically, back then, I was into the "macho flash" school of communication: "this debugger is the best thing since breakpoint debugging, it will solve problems you didn't even think could be solved, etc". I should have just done a very simple demo walkthrough of printf("%d\n", gettimeofday()).
Auxon and Skubeedooo, yeah, it was a lot about marketing.
Jeff Mahoney: I agree, Hindsight looks much more powerful. But Hindsight is also more resource-intensive: they have to simulate a whole CPU.
MenTaLguY: I would be happy to chat with anybody who wants to do a revival. The mec@shout.net contact address still works.
And the whatever-happened-to-mec line: I worked for Cygnus/Red Hat for several years on gdb. My current job is with Google.
I live in New York City within walking distance of a 400 foot hill. Not a 400 foot building -- a hill!
We have several boats that hold 6,000 people at a time, too. I think that in 6 hours, we could carry about 200,000 people. From the dock, it's a 15 minute walk (uphill) to 100 feet of vertical elevation.
That's not many people by New York standards. We could use more boats.
Guess where I live. Hint: it's the borough that you usually overlook.
And I live in New York City, which gives me a little appreciation for death and disaster.
With respect, I think you've been making some of your points badly. You've been saying "10,000 people died, but that's not important, because lots more people die from other things." Try this instead: "20,000 people died, and millions more are homeless, and millions more are at increased risk from disease. And that's an epic tragedy. But look at this: another 15,000 people are dying every day from protein-energy malnutrition. Reputable Link. That's a big problem, too."
That is, if you want to offer a radical critique, it helps to start by acknowledging the validity of the conventional wisdom, and then going beyond it.
Errr, I'm preaching at you. Anyways, that's my two cents and my soapbox.
Have you ever slept rough? I haven't, but I've come close enough to appreciate how bad it is.
It's untold damage to water systems. That story hasn't come out yet.
As to why it matters: these is a humans vs nature story. We humans use engineering to defend ourselves against the depredations of nature like this. When I read these stories, I think: if the people in Indonesia and Thailand and Myanmar and Bangladesh and India had better engineering, fewer lives would be lost and ruined.
The Richter scale is logarithmic in energy; each point represents approximately 31-fold increase in energy. 1,900 megatons of TNT corresponds to your Richter 8.6 earthquake plus another 0.7, or Richter 9.3.
Of course, the energy release of an asteroid strike may have a very different effect than the energy release of an earthquake. And everybody on earth will have decades of warning (for this asteroid at least).
I lived through a Richter 7.0 earthquake. The building I was in stood up, but it was condemned afterwards.
Well, yeah. The function of a company is not to make a good product -- the function of a company is to arrange so that people who use the good product pay for the good product, and then split some of that money with engineering.
I think the numbers you're using include infant deaths. The huge decrease in infant deaths affects the final number a great deal without extending anyone's ability to go deeper into "old age".
(Rats, now I wish I hadn't posted the grandparent article as AC).
Yes and no. You are right: decreases in infant and child mortality brought up the average a lot, but didn't do anything for the life expectancy of people who survived childhood. "average life expectancy at age 0" is not relevant to a 20-year-old or a 40-year-old or a 60-year-old.
However, life expectancy for old people has been going up too. Here are some tables:
For example: American white male, age 60. Life expectancy in 1900: 14.35 years. Life expectancy in 2000: 20.0 years. That is a 39% increase in one century.
That's not as pronounced as "American white male, age 0", which improved from 48.23 years to 74.8 years, an increase of 55% in one century.
Again limited to the United States. The second table is a nice PDF, "Life Expectancy at birth and 65 years of age by sex and race, 1900-2000".
The earliest figures for 65 year old life expectancy are from 1950. Let's grab one: "all races, both sexes, 65 years old". 1950 expectancy: 13.9 years. 2000 expectancy: 18.0 years. That's a 29% increase in half a century.
I don't have any figures for maximum life span, which would be hard to measure. I agree with you that it doesn't look like we're going to get to 180 just by curing all forms of cancer and stenting up our hearts.
What I'm trying to say here is that "lifespan extension" appears to me to be something of an illusion."
Lifespan extension is real.
But at this rate, in 2050 the life expectancy of an American 60-year-old will be about 23 years. (29% increase from 18.0 years). That is nowhere near "live long enough to live forever".
Personally, I am hoping for "live long enough for a revolution in medicine" or "live long enough for workable cryonic suspension" or "live long enough for uploading".
Click on the "Multimedia: Interactive Feature -- Complete Results" for a nice PDF of all the questions and answers. Scroll down to page 5 of the PDF, and look at questions 8 and 9.
(IF ANSWERED "GEORGE W. BUSH" to Q.5, ASK:) 8. Would you describe your support of George W. Bush as strongly favoring him, or do you like him but with reservations, or do you support him because you dislike the other candidates?
strongly favor 70%, like with reservations 22%, dislike others 8%, dk/na 1%
(IF ANSWERED "JOHN KERRY" to Q.5, ASK:) 9. Would you describe your support of John Kerry as strongly favoring him, or do you like him but with reservations, or do you support him because you dislike the other candidates?
strongly favor 48%, like with reservations 26%, dislike others 25%, dk/na 1%
So the people who prefer Bush are pretty solid. This might be consistent with your view that lots of Republicans are unhappy with Bush -- they might be unhappy enough that they didn't answer "George W. Bush" to question 5.
But the people who prefer Kerry are not all that strongly behind. 25% of them still say that they prefer Kerry because they dislike the other candidates, compared to 8% for Bush supporters.
Read the actual questions and answers; there are lots of interesting tid-bits. For instance, in Question 81, 40% of the people polled say that the believe that George Bush did not legitimately win the election -- that's a surprisingly large number of people who don't trust the system.
And again in the 2000 election, 29% of respondents said that they voted for Gore, and 35% said that they voted for Bush. Considering that the actual popular vote was much closer than that, it means (a) some people are lying (some people like to lie and claim they voted for the winner) or (b) the sample of this poll is skewed towards Bush, perhaps by the trendy "cell phone effect".
I used to wonder why political candidates paid their own pollsters, but once you start digging into the polls, you can see it's a lot more useful and interesting than just "X% Bush, Y% Kerry".
The United States is divided into 50 states. The legislature of the United States has two chambers.
The upper chamber, the Senate, has 2 senators from each state. Senators serve for 6 years and their terms are staggered, so except for death or resignation or something, each election at most one Senator is elected from a state. Everybody in the state votes, and whoever gets the most votes in the whole state wins the Senate seat.
The lower chamber, the House of Representatives, has 435 Representatives. States with more people get more Representatives. Texas is a populous state; it has 32 Representatives.
Representatives are elected by districts. The state of Texas is divided into 32 districts. In each district, there is a separate vote for Representative, and whoever gets the most votes in that district is the Congressional Representative for that district.
Note my use of the passive voice above: "is divided into 32 districts". Well, someone does the dividing.
Texas has 12.9 registered voters (from the Texas Secretary of State web site). I can't find a breakdown by party, but I found some random guy on the web claiming that the breakdown is about 60% Republican Party, 40% Democratic Party. That would be about 8 million registered Republicans and 5 million registered Democrats.
So if the 32 districts have exactly the same composition as the state as a whole, then the Republicans would win by approximately 60% in each district, and the Representatives from Texas would be 32 Republicans and 0 Democrats.
But slice up the state another way: make 22 districts that are slightly tilted to Democrats with 225,000 registered Democrats and 180,000 registered Republicans per district. Then make 10 more districts that are pure Republican, with 405,000 registered Republicans and no Democrats per district.
The Democrats will tend to win by narrow margins in a lot of districts, and the Republicans will win by huge margins in a small number of districts. The Representatives from Texas would be 10 Democrats and 22 Republicans.
That's the heart of the problem. By changing the lines between districts, the people who draw the lines can change the likely number of Representatives in the whole state's delegation.
A typical gerrymander move is to create numerous districts where the favored party has a slim but safe margin, and just a few districts where the unfavored party has a very lopsided concentration. By gerrymandering, a minority party can hold onto a lot more seats than it really deserves -- or a majority party can push a minority party (literally) right off the map.
If you still have trouble seeing this, here's a toy example. Consider a state with 100 R's and 50 D's, and five districts. The R's would like to see this:
District 1: 20 R and 10 D
District 2: 20 R and 10 D
District 3: 20 R and 10 D
District 4: 20 R and 10 D
District 5: 20 R and 10 D
Election result: 5 R wins, 0 D wins
The D's would like to see this:
District 1: 14 R and 16 D
District 2: 14 R and 16 D
District 3: 14 R and 16 D
District 4: 29 R and 1 D
District 5: 29 R and 1 D
Election result: 2 R wins, 3 D wins
That's why political parties fight so bitterly over the procedural rules for who gets to draw the lines. In most states, the lines are drawn by the state legislature, which is made up of (surprise) elected officials from major political parties.
Gerrymandering is not an exact science. Just because voter is registered does not mean that they will vote, and voters are not bound to vote for the candidate of the party that they are registered in. But, like insurance companies, political parties can rely on statistical averages of behavior over millions of people, based on decades of past experience.
Similarly, it's hard to devise an objective test for whether a given set of district lines is gerrymandered or not
And who put the current draft registration system in place? Republican President Jimmy Carter. The bill was filibustered in the Senate by Democratic Senator Mark Hatfield of Oregon, but the Senate overrode the veto.
Oh wait. Reality check. Carter was a Democrat, and the opposition and filibuster were conducted by Republicans.
That's how elections work. It's in the Democrat's interests to muck-rake any Republican crimes, and it's in the Republican's interest to muck-rake any Democrat crimes.
I agree with you that there is a market for somebody who can synthesize all this into an overview of American election flaws. And I don't think Carter is that person. He has a lot of integrity, but he is a Democrat partisan.
(Two other Carter bits: in 1980 Carter himself made a concession speech before the polls closed on the west coast, depressing Democratic turnout. And while President, Carter re-instituted draft registration, which is back in the air again).
Slashdot Mirror
Indeed it is mandatory, because SCO probably claims ownership of a trademark on "FSCK".
Just whack Rodney with a cockatrice corpse.
You realize that you and I are part of the media right now, don't you?
In the example I cited, the FBI didn't make its opposition research available to the voters. Hundreds of background files somehow, mysteriously, ended up in an office in the White House.
So ... you want the backgrounds of potential leaders of this country to go unchecked?
Unchecked by whom?
I absolutely DO NOT WANT the White House to be secretly checking FBI background files on potential leaders.
I want media, political organizations, and voters -- on all sides -- to do the checking.
It's not about whether the FBI monitors little old me. (Privacy advocates, hold on for a minute -- I care about little old me, too, but Gravis Zero doesn't).
It's about whether the FBI monitors popular political leaders and whether those FBI files somehow end up in the White House from time to time.
Dude, leave me alone, I'll pull an all-nighter, and you will have an *awesome* new light bulb in the morning. Or by Friday, at the latest.
(The downside: after the software engineer changes your light bulb, you discover that your washing machine doesn't work quite right any more).
If they do something like SCO did, then you can start up the hating.
Sun expands Unix deal with SCO
Sun paid about $10 million to SCO and received warrants to buy 210,000 shares of SCOX as part of the deal.
I do agree with you that Sun has also done some awfully nice things for the open-source world. Sun is a friend when it's in Sun's interest (buying StarOffice and releasing the source under GPL) and a foe when that is in Sun's interest (funding SCO).
Think about if your post offices kept copies of all the mail you received, even after you had thrown it away.
They do. At least, they keep an image of the outside of the envelope.
Uh, how about that other large pool of citizens whose votes determine foreign policy, but who haven't served in the military?
Don't get me wrong -- I do appreciate your insight about matching up voting rights with responsibility with the consequences.
Cue up the Starship Troopers argument!
Computers also don't often:
4) black out at the helm while the backup computer is AWOL.
Allision of Staten Island Ferry Andrew J. Barberi
Pilot Richard Smith blacked out at the helm. Captain Michael Gansas was absent -- his duty station was in the forward wheelhouse to backup Smith, but he wasn't there. As a result of this double human failure, 11 people died.
Humans do fail, and so do computers. I don't think Smith's weakness was excusable but I have sympathy for him. Humans do fail, excusably or inexcusably. That's why the DOT pays a second human pilot to be in the wheelhouse during docking, but Gansas wasn't even there.
I would love for my wristwatch to pop up an alert telling me the optimal time to leave my office.
However, I think the MTA would take a dim view of:
-- Lots more people waiting to the literal last minute to catch a train, and then delaying 30 seconds more, and then running down the stairs.
-- Increased train dwell times from people mis-optimizing their platform arrival experience and forcing the doors open longer.
-- People noticing when trains are significantly late, or even just insignificantly late, and bitching about it, making life unpleasant for the MTA overlords.
Unfortunately, it's human nature for people to be dumbasses, and it's also human nature for large organizations to protect themselves against both dumbasses and smart complainers by hoarding information.
Also I think you've overestimated the positive benefits. If you want to be sure of catching your train, arrive 2 minutes early. I don't know about your job, but with my job, I actually do a lot of thinking away from my desk, and I use those several minutes per day for offline thinking.
A last thought: maybe the customers could do this P2P style with little or no co-operation from MTA. Imagine lots of people carrying PDA's that know where they are in the subway system and can upload real-time data streams to a commercial site. (That would be a badass PDA, and the app would have to run mostly in the background). The commercial site provides the information to your PDA from current riders. You either have to pay money, or watch ads, or leave your own PDA operational during your trip, or some combination of those things.
I've lost about $80 in cash theft in my life.
I've lost $0 in credit card theft, but dealing with the attempts has cost me quite a bit more in time: more than 100 times as much, in time and lost income, than the $80 cash loss.
Cash means more privacy, fewer entries to reconcile, and less paper to shred.
Wow, what to say?!
First, it was kind of silly to name the program the same as my user name, but I never found a better name for it than "my trace-and-replay debugger".
My original plan was to write this for Solaris and sell it, hence the insistence on tracing and replaying without modifying the target program or the operating system, and that's why the replay controller messes with gdb's mind, so that it can work with a stock gdb rather than needing gdb extensions.
I developed a Linux version first because it was so cheap and simple to throw Slackware on my computer. And, well, it turns out that I'll probably never need a Solaris version, because Linux sure has become big enough and rewarding enough for me.
Versions 0.1, 0.2, and 0.3 took 15 months of full-time work, living on my savings and writing code on a little Linux box.
After version 0.3, several bad things happened:
Technical butterfly-chasing: the tracer needs to know about all possible ioctl calls that the target program makes, and Linux was adding and changing ioctls faster than I could check the patch diffs to update them. (That's how I came to write the kernel change summaries for a while). The obvious solution is to trace about 20 common ioctl's and throw all the rest in a big worst-case box that says "this ioctl might touch all of memory". One of the problems of working alone: nobody else around to notice the obvious solution.
Moving up the management chain: anybody who runs a one-person operation knows about this. It's one thing to write a proof-of-concept program. It's another thing to push it out, start a community, market it, manage all the communication with users and co-developers. I failed at that.
Fade-out: after the proof-of-concept worked, I noticed I'd spent 15 months full time, and I did make the milestone of seeing test programs run. I lost some interest and went and did something else.
Version 0.3 still has one good use: to help defend against anybody else that files a patent for technology like this. I released in November 1995.
Some responses:
Zogger and Animats, that's exactly the use case, the user in the field runs the tracer, then mails a big log back to the developer. This gets very useful when the user has unique resources that the developer doesn't, for a program like a network server. I don't know much about dtrace, but I think dtrace is just more comprehensive kernel reporting information, not fundamentally "video-taping the user process".
Mebane, I think the answer is: in 1995, I sucked at explaining things to people. Specifically, back then, I was into the "macho flash" school of communication: "this debugger is the best thing since breakpoint debugging, it will solve problems you didn't even think could be solved, etc". I should have just done a very simple demo walkthrough of printf("%d\n", gettimeofday()).
Auxon and Skubeedooo, yeah, it was a lot about marketing.
Jeff Mahoney: I agree, Hindsight looks much more powerful. But Hindsight is also more resource-intensive: they have to simulate a whole CPU.
MenTaLguY: I would be happy to chat with anybody who wants to do a revival. The mec@shout.net contact address still works.
And the whatever-happened-to-mec line: I worked for Cygnus/Red Hat for several years on gdb. My current job is with Google.
Almost, almost ...
In Soviet Matrix, Beowulf cluster imagines YOU!
The guy who started del.icio.us said that he bought the icio.us domain because it had lots of subdomain possibilities.
mal.icio.us is currently empty, but it would be cool to see something there!
I live in New York City within walking distance of a 400 foot hill. Not a 400 foot building -- a hill!
We have several boats that hold 6,000 people at a time, too. I think that in 6 hours, we could carry about 200,000 people. From the dock, it's a 15 minute walk (uphill) to 100 feet of vertical elevation.
That's not many people by New York standards. We could use more boats.
Guess where I live. Hint: it's the borough that you usually overlook.
Since you ask, yes, I'm a software engineer.
And I live in New York City, which gives me a little appreciation for death and disaster.
With respect, I think you've been making some of your points badly. You've been saying "10,000 people died, but that's not important, because lots more people die from other things." Try this instead: "20,000 people died, and millions more are homeless, and millions more are at increased risk from disease. And that's an epic tragedy. But look at this: another 15,000 people are dying every day from protein-energy malnutrition. Reputable Link. That's a big problem, too."
That is, if you want to offer a radical critique, it helps to start by acknowledging the validity of the conventional wisdom, and then going beyond it.
Errr, I'm preaching at you. Anyways, that's my two cents and my soapbox.
It's more than 14,000 deaths.
It's millions of people homeless.
Have you ever slept rough? I haven't, but I've come close enough to appreciate how bad it is.
It's untold damage to water systems. That story hasn't come out yet.
As to why it matters: these is a humans vs nature story. We humans use engineering to defend ourselves against the depredations of nature like this. When I read these stories, I think: if the people in Indonesia and Thailand and Myanmar and Bangladesh and India had better engineering, fewer lives would be lost and ruined.
Hiroshima = 15 kilotons
So 10,000 hiroshima bombs = 150 megatons.
The Richter scale is logarithmic in energy; each point represents approximately 31-fold increase in energy. 1,900 megatons of TNT corresponds to your Richter 8.6 earthquake plus another 0.7, or Richter 9.3.
That would make it #2 on the largest earthquakes of the 20th century.
Of course, the energy release of an asteroid strike may have a very different effect than the energy release of an earthquake. And everybody on earth will have decades of warning (for this asteroid at least).
I lived through a Richter 7.0 earthquake. The building I was in stood up, but it was condemned afterwards.
Well, yeah. The function of a company is not to make a good product -- the function of a company is to arrange so that people who use the good product pay for the good product, and then split some of that money with engineering.
I think the numbers you're using include infant deaths. The huge decrease in infant deaths affects the final number a great deal without extending anyone's ability to go deeper into "old age".
(Rats, now I wish I hadn't posted the grandparent article as AC).
Yes and no. You are right: decreases in infant and child mortality brought up the average a lot, but didn't do anything for the life expectancy of people who survived childhood. "average life expectancy at age 0" is not relevant to a 20-year-old or a 40-year-old or a 60-year-old.
However, life expectancy for old people has been going up too. Here are some tables:
Life Expectancy by Age, 1850-2001
For example: American white male, age 60. Life expectancy in 1900: 14.35 years. Life expectancy in 2000: 20.0 years. That is a 39% increase in one century.
That's not as pronounced as "American white male, age 0", which improved from 48.23 years to 74.8 years, an increase of 55% in one century.
NCHS - FASTATS - Life Expectancy
Again limited to the United States. The second table is a nice PDF, "Life Expectancy at birth and 65 years of age by sex and race, 1900-2000".
The earliest figures for 65 year old life expectancy are from 1950. Let's grab one: "all races, both sexes, 65 years old". 1950 expectancy: 13.9 years. 2000 expectancy: 18.0 years. That's a 29% increase in half a century.
I don't have any figures for maximum life span, which would be hard to measure. I agree with you that it doesn't look like we're going to get to 180 just by curing all forms of cancer and stenting up our hearts.
What I'm trying to say here is that "lifespan extension" appears to me to be something of an illusion."
Lifespan extension is real.
But at this rate, in 2050 the life expectancy of an American 60-year-old will be about 23 years. (29% increase from 18.0 years). That is nowhere near "live long enough to live forever".
Personally, I am hoping for "live long enough for a revolution in medicine" or "live long enough for workable cryonic suspension" or "live long enough for uploading".
But I can't buy that a majority of Democrats aren't happy with Kerry
Here is a New York Times / CBS poll with some interesting data for you.
NYT Article
Click on the "Multimedia: Interactive Feature -- Complete Results" for a nice PDF of all the questions and answers. Scroll down to page 5 of the PDF, and look at questions 8 and 9.
(IF ANSWERED "GEORGE W. BUSH" to Q.5, ASK:)
8. Would you describe your support of George W. Bush as strongly favoring him, or do you like him but with reservations, or do you support him because you dislike the other candidates?
strongly favor 70%, like with reservations 22%, dislike others 8%, dk/na 1%
(IF ANSWERED "JOHN KERRY" to Q.5, ASK:)
9. Would you describe your support of John Kerry as strongly favoring him, or do you like him but with reservations, or do you support him because you dislike the other candidates?
strongly favor 48%, like with reservations 26%, dislike others 25%, dk/na 1%
So the people who prefer Bush are pretty solid. This might be consistent with your view that lots of Republicans are unhappy with Bush -- they might be unhappy enough that they didn't answer "George W. Bush" to question 5.
But the people who prefer Kerry are not all that strongly behind. 25% of them still say that they prefer Kerry because they dislike the other candidates, compared to 8% for Bush supporters.
Read the actual questions and answers; there are lots of interesting tid-bits. For instance, in Question 81, 40% of the people polled say that the believe that George Bush did not legitimately win the election -- that's a surprisingly large number of people who don't trust the system.
And again in the 2000 election, 29% of respondents said that they voted for Gore, and 35% said that they voted for Bush. Considering that the actual popular vote was much closer than that, it means (a) some people are lying (some people like to lie and claim they voted for the winner) or (b) the sample of this poll is skewed towards Bush, perhaps by the trendy "cell phone effect".
I used to wonder why political candidates paid their own pollsters, but once you start digging into the polls, you can see it's a lot more useful and interesting than just "X% Bush, Y% Kerry".
Sure, here's how it works.
The United States is divided into 50 states.
The legislature of the United States has two chambers.
The upper chamber, the Senate, has 2 senators from each state. Senators serve for 6 years and their terms are staggered, so except for death or resignation or something, each election at most one Senator is elected from a state. Everybody in the state votes, and whoever gets the most votes in the whole state wins the Senate seat.
The lower chamber, the House of Representatives, has 435 Representatives. States with more people get more Representatives. Texas is a populous state; it has 32 Representatives.
Representatives are elected by districts. The state of Texas is divided into 32 districts. In each district, there is a separate vote for Representative, and whoever gets the most votes in that district is the Congressional Representative for that district.
Note my use of the passive voice above: "is divided into 32 districts". Well, someone does the dividing.
Texas has 12.9 registered voters (from the Texas Secretary of State web site). I can't find a breakdown by party, but I found some random guy on the web claiming that the breakdown is about 60% Republican Party, 40% Democratic Party. That would be about 8 million registered Republicans and 5 million registered Democrats.
So if the 32 districts have exactly the same composition as the state as a whole, then the Republicans would win by approximately 60% in each district, and the Representatives from Texas would be 32 Republicans and 0 Democrats.
But slice up the state another way: make 22 districts that are slightly tilted to Democrats with 225,000 registered Democrats and 180,000 registered Republicans per district. Then make 10 more districts that are pure Republican, with 405,000 registered Republicans and no Democrats per district.
The Democrats will tend to win by narrow margins in a lot of districts, and the Republicans will win by huge margins in a small number of districts. The Representatives from Texas would be 10 Democrats and 22 Republicans.
That's the heart of the problem. By changing the lines between districts, the people who draw the lines can change the likely number of Representatives in the whole state's delegation.
A typical gerrymander move is to create numerous districts where the favored party has a slim but safe margin, and just a few districts where the unfavored party has a very lopsided concentration. By gerrymandering, a minority party can hold onto a lot more seats than it really deserves -- or a majority party can push a minority party (literally) right off the map.
If you still have trouble seeing this, here's a toy example. Consider a state with 100 R's and 50 D's, and five districts. The R's would like to see this:
District 1: 20 R and 10 D
District 2: 20 R and 10 D
District 3: 20 R and 10 D
District 4: 20 R and 10 D
District 5: 20 R and 10 D
Election result: 5 R wins, 0 D wins
The D's would like to see this:
District 1: 14 R and 16 D
District 2: 14 R and 16 D
District 3: 14 R and 16 D
District 4: 29 R and 1 D
District 5: 29 R and 1 D
Election result: 2 R wins, 3 D wins
That's why political parties fight so bitterly over the procedural rules for who gets to draw the lines. In most states, the lines are drawn by the state legislature, which is made up of (surprise) elected officials from major political parties.
Gerrymandering is not an exact science. Just because voter is registered does not mean that they will vote, and voters are not bound to vote for the candidate of the party that they are registered in. But, like insurance companies, political parties can rely on statistical averages of behavior over millions of people, based on decades of past experience.
Similarly, it's hard to devise an objective test for whether a given set of district lines is gerrymandered or not
And who put the current draft registration system in place? Republican President Jimmy Carter. The bill was filibustered in the Senate by Democratic Senator Mark Hatfield of Oregon, but the Senate overrode the veto.
Oh wait. Reality check. Carter was a Democrat, and the opposition and filibuster were conducted by Republicans.
That's how elections work. It's in the Democrat's interests to muck-rake any Republican crimes, and it's in the Republican's interest to muck-rake any Democrat crimes.
I agree with you that there is a market for somebody who can synthesize all this into an overview of American election flaws. And I don't think Carter is that person. He has a lot of integrity, but he is a Democrat partisan.
(Two other Carter bits: in 1980 Carter himself made a concession speech before the polls closed on the west coast, depressing Democratic turnout. And while President, Carter re-instituted draft registration, which is back in the air again).