Saturday, 31 August, 2002

Building a Linux From Scratch System

In my July 24, 2001 entry I mentioned Linux From Scratch, and that I was planning on building such a system.  I finally cobbled together another computer (that's three I have running here now, not counting Debra's machine) that I can play with.  The "new" computer is my old Dell Pentium 200 MHz system with 64 MB of RAM.  It's the system I used to write the Kylix book and do my first experiments with Linux.  I took the drives, monitor, and keyboard from the Dell to put into the new system, leaving me with a Dell case, motherboard, RAM, and power supply.  Last night I cannibalized an old 486 system for a few parts, located the Dell's original video card, and threw in a couple of old hard drives (3 GB and 8 GB).  SuSE Linux installed without a hitch, and I now have a minimal SuSE Linux system running in a 1 GB partition.  Following the instructions on the Linux From Scratch site, I've created a 2 GB partition for my LFS system.  I don't know yet what I'll do with the 8 GB drive.  I'll decide once the LFS system is up and running.  I start the LFS installation tomorrow, and will let you know how it goes.

Thursday, 29 August, 2002

Conference Bike

It's not very practical, but it's kinda cool.  Check out the Conference Bike.  It's a tricycle with seats for 7 people, all of whom can pedal or not as they choose.  Before you get any ideas, it can't go very fast.  Their FAQ page says the fastest they've had it going was 35 km per hour (about 22 MPH) on a long downhill.  Under normal riding on a flat, expect about 10 KPH (6 MPH), or 15 KPH if you're really pushing it.  The bike is too heavy (almost 450 lbs) and the aerodynamic cross-section is way too big for it to be any kind of speed machine.  Nonetheless, it'd be great fun riding the thing around an amusement park, or cruising the boardwalk at some vacation spot.  If you have a large family and live in a quiet neighborhood, it'd be an interesting way to spend some "quality time" together.

What distinguishes a religion from a cult?  I don't remember how the subject came up yesterday, but my friend John's immediate answer was "The number of followers."  The six of us talked around the topic all the way back to the office after lunch, and that was still the most satisfying answer, as far as I'm concerned.  Merriam-Webster's definition of the word cult isn't much help, and the Defining Terms article from the Deception in the Church page, although interesting, just adds more fuel to the fire.  Popular usage of the term cult is all about what's considered mainstream.

No, there really isn't a point to all this pondering, beyond noting that one person's (or group's) cult may be another's religion.  Definitions are such slippery things sometimes.

Monday, 26 August, 2002

Fat People Sue Fast Food Companies

I must have been living under a rock last month when this story came out.  I heard the story on Marketplace when I was driving home this evening, and then searched the web to find the Fox News article.  I mentioned this as a possibility in my January 22, 2002 diary entry, but it still bothers me that it's come to pass.

A New York City lawyer has filed suit against the four big fast-food corporations, saying their fatty foods are responsible for his client’s obesity and related health problems.

The lead plaintiff, Caesar Barber, is a 56-year-old maintenance worker who ate at fast food restaurants four or five times a week, and blames his diet for his many health problems:  obesity, diabetes, high blood pressure, high cholesterol, and two heart attacks.  He's the only complainant named in the suit right now, but two others will be filing soon:  a 57-year-old retired nurse who says eating fast food twice a week since 1975 caused her to go from a size 6 to a size 18, and a 59-year-old man who says that his habit of eating a pound of fries a week gave him high blood pressure, diabetes,  and made him obese.  "In 1993," the article says, "he passed out and had to be rushed to the emergency room because of the medical problems caused by his diet."

"I trace it all back to the high fat, grease and salt, all back to McDonald’s, Wendy’s, Burger King – there was no fast food I didn't eat, and I ate it more often than not because I was single, it was quick and I’m not a very good cook," Barber said.  "It was a necessity, and I think it was killing me, my doctor said it was killing me, and I don't want to die."

I'll bet his doctor said something about exercise, too.  He's going to sue the fast food industry because he's too lazy to learn how to cook?  This is ridiculous!  Talk about living under a rock!  How can these people try to hold the food industry responsible for their own bad choices?  The attorney in the case claims that the aim of the legal action is to force fast food restaurants to offer healthier choices, and to obtain federal legislation that would require warning labels on fast food similar to those on tobacco products.  Yeah, right!  If they don't listen to their doctors saying "that stuff's gonna kill you," it's damned unlikely that these fat asses would give a second thought to a warning label.  Not that warning labels would save the fast food industry even if people did read them.  We've seen how effective those labels were in protecting the tobacco industry.

I doubt that this particular suit will go anywhere, as public sentiment isn't quite yet on the side of the fat asses.  I think similar suits will be filed, though, with increasing frequency.  But in 10 or 20 years, when more of the population (and most of the judges and lawyers) are fat assed couch potatoes facing the consequences of their lifetime of sedentary self indulgence, they'll extort billions of dollars from the fast food industry in the same way that the States have robbed the tobacco industry.

Personal responsibility?  Ha!  Who needs it?

Sunday, 25 August, 2002

Century Training Progress

The century training continues.  This week was a bit different because my road bike was in the shop for a couple of days.  So I got to ride the mountain bike instead.  Other than that, just training as usual.  Monday I was supposed to do 12 miles easy, but I decided to ride to the office with a co-worker, so I ended up doing 13.5 in the morning and in the evening.  Since last Sunday was kind of an easy day, I did hill sprints on Monday.  That is, I rode slowly along the flats, coasted on the downhills, and sprinted like mad up the hills.  Runners call it "fartlek."  Cyclists call it "unstructured interval training."  I'm feeling very good, and continue to be impressed by the training program.

This week's mileage:

Monday 27 varied
Tuesday 17.7 pace
Wednesday 18 brisk
Thursday rest
Friday 18 pace
Saturday 59.2 pace
Sunday 22 pace
Total 163.4


Saturday, 24 August, 2002

Extreme Bike Rides

It takes discipline and hard work to train for and ride 100 miles on a bicycle.  There's no doubt about that.  But in cycling circles, "simply everybody" has done a century.  In my two years of road riding, I've met many people who have moved from overweight couch potato to century finisher in relatively short periods of time. Some in less than six months.  Accomplishment that it is, I'll have to do something a bit more intense in order to remain interested.  Just as with running, where "simply everybody" has run a marathon, but not so many have done Pike's Peak.

While I was researching training programs, I ran across two very interesting groups.  Members of the Ultra Marathon Cycling Association specialize in endurance races:  double centuries (200+ miles), 12-hour and 24-hour races, and the annual Race Across America (RAAM).  I won't be riding 3,000 miles in 8 days any time soon, but a double century looks interesting.  Perhaps next summer.

The other group is the Perimeter Bicycling Association of America. What these folks do is intriguing—they ride around things: cities, lakes, counties, countries, etc. One person has logged almost 65,000 miles in perimeter rides, covering 9 countries, 26 counties, 31 lakes, and 5 islands.  This is something I can sink my teeth into. I can think of plenty of things to "ride around," right here in Central Texas.  I'll need to improve my fitness level a bit, but riding around the city of Austin, or the perimeter of Lake Travis don't look like unreasonable goals to start.

The August 19 issue of eWeek has a group of articles about spam.  One article describes the problem, and provides some interesting numbers about how spam has grown over the last year, and what spammers are selling.  Other articles discuss current filtering techniques, and some proposed legislation.  What strikes me about these articles is that almost all of the "experts" agree that filtering is at best temporarily effective, and that legislation likely will be wholly ineffective.  And yet, those same "experts" continue to support legislation.  Huh?

Just how effective do you expect anti-spam legislation to be when the very same legislators who will be passing the laws will use unsolicited bulk email as a tool to get re-elected?  If you doubt that, take a look at this article from The Mercury News.  In related news, the Federal Election Commission decided that it's okay for political ads transmitted over SMS (short messaging service) to forego transmitting disclosure information.  To be fair, disclosure exemptions are common practice for media that are limited to small numbers of characters.  Still, expect your mobile phones and perhaps your text pagers to be flooded with political spam in the next few months.  I'm sure your newly elected legislators will be happy to pass an anti-spam bill after they see how effective those campaigns are.

I've mentioned the Internet Mail Consortium before.  Today I found two reports on their site:  Unsolicited Bulk Email: Definitions and Problems, and Unsolicited Bulk Email:  Mechanisms for Control.  Both reports were written in 1997, and the "Mechanisms for Control" report was updated in May of 1998.  The articles do a good job of defining the problem and identifying in broad terms the possible solutions.  One thing of note is that this is the only place I've seen serious discussion of my "trusted server" idea.  They call it "First-hop Accountability."  I find nothing on their site that takes this idea any further, nor any real discussion of the spam problem in general, other than support of legislation.  I'm disappointed, but not terribly surprised, I guess, by the IMC's ineffectiveness.  Their main web page says:

The Internet Mail Consortium is the only international organization focused on cooperatively managing and promoting the rapidly-expanding world of electronic mail on the Internet. The goals of the IMC include greatly expanding the role of mail on the Internet into areas such as commerce and entertainment, advancing new Internet mail technologies, and making it easier for all Internet users, particularly novices, to get the most out of this growing communications medium.

In truth, I think the member organizations are members only to protect their own interests, and ensure that they're kept abreast of any proposed changes.  That they haven't actually done anything to help combat the spam problem shows me that the IMC is just another group of industry "leaders" who aren't at all interested in solving anything.

Finally, I wonder how receptive Brightmail and other anti-spam software providers would be to an effective solution to the spam problem.  If I were a conspiracy theorist, I might even accuse these companies of supporting spammers.  But, no, they wouldn't do that.  Would they?

Thursday, 22 August, 2002

Pike's Peak Marathon

20 years ago today, I ran the Pike's Peak Marathon.  That is by far the most physically challenging thing I have ever done.  Starting in Manitou Springs, you travel 13.9 miles up the mountain; first on a pretty decent 4-wheel-drive road, and then a mule trail that winds its way across the mountain and up.  The elevation of the starting line is about 7,500 feet.  The peak itself tops out at 14,110.  The last three miles of the climb are above timberline, with an average rise of 1,000 feet per mile.  It took me three hours and 45 minutes to get to the top of the mountain.  The 14.1 miles back down the mountain weren't a whole lot easier.  I slipped and fell several times on the way back down the first three miles, and after that the steep grades made me backpedal to keep from going too fast and falling over.  I finished the last 10 miles in a mental fog.  Even 20 years ago, I remembered very little of the descent.  Other than my falls, I have two very distinct memories:  the woman with whom I shared most of the last 5 miles, and the spectators all along the last mile to the finish yelling "It's just around the next corner!"  By the time I finally turned a corner and saw the finish line, I wasn't entirely sure it was real.  I was completely exhausted after a little over 6 hours of running.

Where were you 20 years ago?

Oh, and in case you're wondering, the answer is 2.5 pounds per year.  I'm trying to get that down to about 1.5, but it's a tough go.

Monday, 19 August, 2002

Drawbacks to Full Text Searching

I think I've mentioned before the limitations of full text search.  Today I got a couple of great examples.  I'm doing some research on watershed management, flood control, how agencies manage the floodgates in our many dams.  My Google search term today was "watershed modeling floodgate".  The first thing to catch my eye was the second sponsored link from Fox Talent says "Are you model material?".  Apparently, Fox Talent pays for every hit on the term "modeling".  The sixth link on the first page of search results was this article from Christianity Today magazine.  Need another?  Last week I got an email from a woman who was given a link to my December 2000 diary page when she searched for "fig toothache."  She was looking for information on a folk remedy for toothache, which involves the sap from a fig tree.  She got my page because early in the month I had mentioned my toothache, and late in the month I mentioned trimming back the fig trees.

It sure would be nice to have a well organized "table of contents" for the Web.  Full text search has its uses, but as the web grows, the signal to noise ratio in Google searches is becoming absurd.

Sunday, 18 August, 2002

Training for a Century

The reason I've been so quiet here recently is that I've started a training program that has me getting up at 5:00 almost every morning to go cycling.  I've set a goal to ride the Waco Wild West Century on September 28, and I'm determined to do it right this time.  The last time I rode 100 miles (two years ago), I wasn't prepared for it and I suffered through the last 25 miles.  My body hurt, and it took me a week to recover.  It was not a pleasant experience.  My "training" for that ride consisted of riding to work and back (26 miles each way) a couple of times a week, and perhaps a long ride of 50 or 60 miles on Saturday.  Sure, I was riding a lot, but I was doing it wrong.

After looking through several books and browsing the Web for relevant articles I settled on this training program from Bicycling magazine (oddly enough, I can't find that thing on the Bicycling web site), and started it last Monday.  I rode 50 miles relatively comfortably last Saturday, so I figured I was fit enough to tackle the second, more aggressive schedule.  A week into the program I'm feeling great—none of the run down feeling or continual sore legs that I remember from the last time.  My legs are a bit weak from yesterday's 52 mile ride in the wind and the hills, so I made today an easy ride rather than the "pace" ride that the schedule calls for.

This week's miles:

Monday 11.5 easy
Tuesday 16.4 pace
Wednesday 20.2 brisk
Thursday rest
Friday 16 pace
Saturday 52.5 pace (hard ride)
Sunday 18 easy
Total 134.5

Granted, I had reached a certain level of fitness prior to starting the program, but in the past I've had difficulty getting beyond that point.  It seems like 50 miles is my "easy" threshold—going beyond that takes a concerted training effort.  So far, so good.  I'll update my progress here each Sunday until the ride.

Friday, 09 August, 2002

Studying Traffic

Traffic is a thorny problem.  I've been running across some some rather interesting traffic research in my hunt for some project related data (looking for electronic maps and traffic count data).  Mostly I try to skip over the research because I don't really have the time to read it, but the Texas A&M University 2002 Urban Mobility Study caught my eye.  If you're at all interested in traffic congestion (and who isn't, if they commute to work), you should take a look.  I found the Issues and Measures page particularly interesting.  The pages linked from that page do a very good job of explaining the current state of the problem, and identifying many of the factors that cause urban congestion and why it's so difficult to solve.  For example, in the discussion of why just adding roads can't solve the problem, there is the answer to the common question "Why don't I see much relief in my travel time [when a road is added or expanded]?"  The answer lies in what's called a "triple convergence:"  travel moves to the new road from other times, other roads, and other modes.  Overall, it's an interesting study, and well worth the read.

If you're looking for road map data, by the way, your best bet is your state's Department of Transportation.  In Texas, the DOT gives the data to the Texas Natural Resources Information System (TNRIS), where you can download files, or obtain them via CD (for a charge).  The files are in GIS formats like DWG and E00, so you'll need a program like AutoCAD (with the mapping module) or ArcInfo to read and display it.  Traffic count information is a bit more difficult to come by because it's not always collected by the state.  Many city and county governments collect traffic counts, and there doesn't appear to be a centralized repository, at least not that I've found in Texas.

Geographic Data Technologies (GDT) sells road map and traffic count data, by the way.  The nice thing about dealing with them is that you can get everything from a single place.  It's expensive, though, if you're looking for data on every state.  Expect to pay upwards of $10,000.00 for nationwide road map information.

Yahoo is running this news story about the effects of contrails (condensation trails formed by jets passing through the air) on local temperature variation.  The long-held belief was that contrails, which often create cirrus clouds that remain for long periods of time, reduced daytime temperatures in local areas and reduced nighttime cooling.  Scientists compared the local temperatures in selected areas during the four days following September 11 (when there was no commercial air traffic) against the averages, and found that there is an indication that the theory is correct:  contrails are causing local temperature variations.  The Yahoo story (grabbed from AP) is a pretty poor summation of the research, by the way.  A Wired story from May 15 is a bit better, but you're probably better off finding the article in the next issue of Nature.  Especially bad was an interview on NPR (I think it was Morning Edition) where somebody characterized the results as providing "irrefutable evidence".  There are plenty of grounds on which to refute the evidence.

Contrails, by the way, are water vapor—a by-product of burning jet fuel.  Contrary to popularly held belief, they are not "exhaust" in the sense of soot or other noxious chemicals.  The cold atmosphere at those altitudes (25,000 to 45,000 feet, usually) won't hold the water vapor, so it condenses to form clouds.  A good place to start learning about contrails is this article over at HowStuffWorks.

Tuesday, 06 August, 2002

Edsger Dijkstra, R.I.P.

Edsger Dijkstra, one of the founders of computer science, passed away today after a long struggle with cancer.  In most programming circles, he's well known for his article Go To Statement Considered Harmful, which first appeared in 1968.  Anybody with a computer science degree, or who has studied algorithms, probably is familiar with his shortest path algorithm.

It's a sad day for computer science, but he does leave us plenty by which to remember him.

Sunday, 04 August, 2002

Hotmail Gets A Billion Spams a Day?

The Houston Chronicle is running a three-part series of articles on spam.  No, not the delicious Hormel treat, but the trash email that we all know and love.   Most of us probably already know what's in the linked article, although I found this paragraph to be incredible:

On a typical day, Hotmail subscribers collectively receive more than 1 billion pieces of junk e-mail. Such spam accounts for 80 percent of messages received -- not including mail blocked by Hotmail's first line of filters. 

80% of delivered mail on Hotmail is spam?  I wonder what percentage of incoming mail gets blocked by the first line of filters.

Hotmail is perhaps an extreme example.  As one of the largest (if not the largest) email provider, it's a perfect target for spammers who use dictionary attacks.  They'll just send mail to "jim@hotmail.com", "jim1@hotmail.com", "jim2@hotmail.com", etc.  Mail to invalid addresses simply falls into the bit bucket.  And emails that are read typically have a fake image link that lets the spammers know that they've found a valid email address.

It's frustrating to watch and listen as people fight with this problem, when there are perfectly reasonable and easily implemented alternatives that will virtually eliminate spam.  But people resist change, and would rather complain about an unpleasant situation than cause themselves some short-term inconvenience in order to solve the problem.  So now I just sigh and shake my head.

Friday, 02 August, 2002

Looking Into Genetic Programming

Most programmers have had some experience with a code generator of one type or another.  In 1984, for example, I and a friend wrote a program that would generate COBOL user interface code, given a description of the screen (in a text file) and the names and types of variables that were being displayed.  It wasn't anything fancy, but it saved us hours of typing very repetitive code.  The Unix tools lex and yacc are compiler creators—given a formal description of a programming language, these tools will create a compiler front-end.  Visual Basic,  Delphi, and other similar development environments also have code-generation features.  They're great time saving devices.

But the real dream is programs that write whole programs, given only a high-level description of a problem.  That's the goal of genetic programming.  My work and personal research interests have combined to lead me down this particular path.  I'll (I hope) have a lot more to say about it in the future.

My friends and co-workers Andrew Reid and Ron Hollis over at Lone Star Handcycling decided to enter the handcycling division of theNew York Marathon.  Andrew, who has been in a wheelchair for about 2-1/2 years, and who recently won the Canadian national championships, sent in his entry a while back.  Ron called today to see about entering.  The guy on the other end of the phone (the race organizer, I assume) refused to let Ron enter because he is able bodied (i.e. he can walk).

Now I'm not one to start waving the Americans with Disabilities Act or other non-discrimination laws in somebody's face, but it seems ironic to me that somebody who supposedly fights against discrimination of the handicapped would turn around and discriminate against somebody because he's able bodied.  It's not like Ron will be using his legs to compete.  He rides a handcycle just like everybody else.  It sure is a strange world.