Saturday, December 31, 2005

Narnia

I gave my 9 year old, 3rd grader son a copy of The Lion, The Witch, and the Wardrobe for Christmas. He has already consumed it. So, perhaps the entire boxed set of books should have been purchased. The books aren't expensive, but they are cheaper all at once rather than one at a time.

Perhaps the book should have been called The Wardrobe, the Witch, and the Lion, as that is the order that they appear in the story. Would that have been written by Lewis S. C.? Clearly, the writer wanted the lion to get top billing, rather than a box.

This particular purchase was inspired by the newly released movie by the same name. It could be pointed out that this isn't the first time the book has been made into a movie, just the first big budget production.

I had never read the books, but had read some other C. S. Lewis works, and had some idea what it might be about. So, though my 9 year old started reading it first, i borrowed it while he slept and finished it off. When he was finished with it, we saw the movie together.

It should be noted that the book isn't that long. In fact, it is about the same length as a screenplay. So, there is no need for the movie to cut anything. And, they didn't. They did add some explicit background at the beginning explaining why the kids were sent away that wasn't in the book. They added an exciting river scene as well. There were a small number of other minor changes.

One change made for the movie was that Aslan, the king lion, is supposed to be frightening as well as good in the book. Yet, no one is obviously scared out of their wits, even when seeing him for the first time. That certainly could have been done in the movie. Perhaps they wanted a true G rating. That's probably why the battle scene has so little violence and no gore at all. Compare that with The Return of the King. This is a choice, probably made for monetary reasons, but welcome for parents of smaller children who want to provide some entertainment for their kids.

As a modern, high budget film, the special effects were mostly excellent. We've come to expect this, and perhaps are spoiled. So, when they use a real dog and just animate the mouth for speaking, it looks so old school. This film has lots of talking animals. And, kids love that sort of thing, for no apparent reason.

The book/movie combination is a great combination. It is an opportunity for parents to read the book to their smaller children before letting them see the movie. For parents of slightly older children, promising to show them the movie after they've read it is a powerful incentive. The combination has special powers to improve reading enjoyment and comprehension, not just for this story, but for stories to come. I've followed this strategy with the first four Harry Potter Movies, when I read the book aloud. This is the first story where my son could reasonably read the book himself.

Thursday, December 29, 2005

What I Don't Like About The Web

Web pages seem to be created by web designers and web developers using local network bandwidth, brand new high end computer hardware, with the latest web browsers that are loaded with every imaginable plug-in. Web pages seem to be never tested in any other environment.

For users at businesses, high speed network connectivity with hardware that is three years old or less seems to be the norm. These users often put up with slow pages, but the web sites often at least work.

One should be able to use Netscape 3.1 on my 1987 vintage Mac II, with my 2400 BAUD external modem, and pages should come up quickly. OK, so the real time kids games won't happen. Also, it doesn't matter that movies would take forever to download. The '87 Mac won't show them anyway, despite the fact that Quicktime was originally written on the 1987 Mac II. A 16.7 MHz 68020 just doesn't have the horsepower for full screen movies, without regard to video acceleration. For that matter, it takes an eon just to decode and display a full screen jpg image. Still, one ought to be able to read text with a few pictures without the thing locking up, requiring a reboot. I should be able to connect to my bank without Java, Javascript or Flash, and get my balance, and make simple transfers. And next year, one should not have to upgrade my computer, operating system, and browsing software just to do it. Netscape 3.1 understands forms, images and styled text. That's where almost all the content on the internet that is at all interesting is still.

OK, maybe 19 years is too much backward compatibility. The web isn't even that old. Perhaps it is amazing that such a machine could ever surf. But ten years isn't out of the question. That puts the bar at 1996. The web existed then. Any browser built back then that can cope with styled text, images and forms should be fine. In particular, Lynx is text only. It knows how to ignore styled text, and images, and handles forms. I can't use it with my bank, because it does not handle Javascript. Yet, my bank ends up using simple forms to actually allow account balances and transfers. Lynx is good enough to show the cnn.com article about newly discovered rings around Uranus. Laugh if you like. This is content.

While we're at backward compatibility, consider PDF documents. One should not post PDF documents that absolutely require the most recent version of the acrobat reader. In particular, I shouldn't have to upgrade to the new version, because zero cost isn't zero cost. Adobe has started inserting adware into their free readers, and on a slow modem, it still takes roughly forever to download the new version. Further, most people on the internet are incompetent to install anything properly. That's right. Most people do not upgrade their computers in any significant way. They either get a friend who is competent to do it for them, or they don't upgrade. Richer people actually buy a new machine when it gets infected with malware, and abandon the old stuff to their basements. I'm not making this up. The upgrade and install burden becomes huge for the competent. Perhaps I should start charging. I don't want a shrubbery. Perhaps pizza.

So, here's an offer. Point me to your web site. I'll test it with Netscape 3.1, 4.7, Netscape 6.x, Lynx, and a couple versions of Firefox. I'll also test with a 33 MHz 486 with 16 MB RAM on a 2400 BAUD modem. Not only that, but I'll test with the '87 Mac II on a 14.4 KBAUD modem. All on a slow modem. My modem is 56K BAUD, but my SBC, sorry, AT&T, (it is hard to keep up with this merger nonsense thing too), phone line is so poor, that 5600 BAUD is not possible. Mostly, 28.8 KBAUD is all i get. I'll tell you in no uncertain terms what is wrong with your site, and all you need to do is send me $100. Cheap.

Sunday, December 18, 2005

State Of Fear

Part of the danger of reading several books at the same time is that more than one can be finished in a short period. So it is now. At random, then, this is a review of Michael Crichton's State of Fear. Only a few of the books in the works are dead trees, and this one was an audio book, checked out of the local library, and consumed, mostly, on the way to or from work.

The plot revolves around global warming, ecoterrorism and related topics. A significant amount of science is bandied about. All of the standard tricks for making a point in science (even when the point is out and out wrong) were used, and many were explained. The author argues from an unpopular point of view, namely, that Global Warming isn't a threat, that it is ill defined and, by the way, evidence for it is difficult to come by. The idea seems to be to make the reader uncomfortable. If that goads the reader into learning some real critical thinking, and perhaps even doing some real research, perhaps this is a good thing. However, The Andromeda Strain, and Jurassic Park were more enjoyable. At least in these books, one could easily tell where the science ends and the fantasy starts.

There was also a break in the suspension of disbelief. One of the characters was supposed to be a mega genius and expert in a variety of fields, including computers. This happens to be an area of expertise for me. The expert just didn't get it. Honey pots and Trojan horses were particularly badly mishandled. These items were not critical to the plot, and proof reading by someone who uses computers for something other than just word processing could have improved the narrative significantly. If the expert wasn't very good with computers after all, then perhaps he wasn't competent in other areas. Without competence, the whole thrust of the novel unravels. Come to think of it, computers weren't handled all that well in Jurassic Park. Perhaps this proof reading thing will have merit in the future as well.

On the positive side, now that the name has been pronounced by an actor, I can, with confidence say that I know how to pronounce Michael Crichton's name. It is pronounced Michael Crichton.

Friday, December 09, 2005

Intelligent Intelligent Design

I've finally gotten through reading an interesting blog on Intelligent Design at Carl Zimmer's The Loom. Carl starts it off by quoting someone name David, and responding. The rest follows.

In it, someone named John A. Davidson explains his Prescribed Evolutionary Hypothesis (PEH), which says that evolution happened at one time, but has stopped. In particular, PEH predicts that no new information is added to genomes, only subtraction is possible. Presumably, this allows for a creator to have put all the information into the DNA of all species, but no further information will ever become available. This is a testable hypothesis, and is the first I've encountered from the Intelligent Design group.

Unfortunately, John A. Davidson states the evidence that would be required for PEH to be falsified, which just isn't how science works. He suggests that one would have to find a mammal species that is new since historic times. Such a species would be considered new if hybrids between it and a related species produces sterile offspring.

With this criteria, one would not expect John A. Davidson, who says he's had that name since 1928, to survive long enough for the evidence he seeks to become available. He asks for hard work. And, if restricted to working with mammals, it could take a very long time.

I haven't read John A. Davidson's published papers. Still, it seems odd that he presents no evidence for offering this hypothesis. Evolution appears in the fossil record, and is shown to take a long time. One wouldn't expect much speciation in the relatively short span of human history, even from a large pool of species. Since absence of evidence is not evidence of absence, it seems clear that something more convincing is required. After all, since the fossil record indicates continuing evolution for the entire history of life on Earth, it seems odd that it would come to a stop in modern times, just for humans to witness. What, exactly, is the proposed mechanism? Well, that there is a Designer, who designed all life, and has stopped working at it. If true, and there is no evidence that it is, what evidence is there that the Designer has stopped designing?

There was a complaint on the blog on why only mammals are considered, to no response. Mammals typically have long generations. For example, humans require nearly 20 years for a generation. Microbes have been shown to adapt to new hostile environments in just 600 generations. But that would require 12,000 years for humans, which makes it unlikely that speciation would happen in the small window made available. Now, many bacteria divide asexually, which has not been shown to generate new genetic information as quickly. And, you can't call the new bacteria a species, and certainly not with the reproductive test proposed. For asexual bacteria, each individual with a mutation can be thought of as a new species. Scientists don't do that, as it isn't a very useful definition. Also, 600 generations yields small changes, and no one would consider the result as a new species.

So, when sea snakes were proposed as a newly recent species, it was ignored.

Dogs and cats were dismissed, too. Dogs were dismissed as wolves. However, wolves may just be feral dogs. Recent evidence suggests a northern Mongolian wild dog species as the origin of domesticated dogs. Is there any hybridization information?

Yet, a paper on gerbils (later referred to as voles) was proposed, and it looks as if it should be a nail in the coffin for this hypothesis. It looks to be time to admit it, and move on to another hypothesis.

Yet the lid should be closed only by showing that there is indeed new useful information produced for the genome of some organism. This wasn't discussed. At least for microbes, every time a microbe becomes resistant to some new antibiotic, one could argue that new useful information is gained. One must, however be careful that it isn't just old information that was turned off, and is now switched back on. So, the critter should be sequenced before the experiment begins, and when it becomes resistant, it should be sequenced again. We might even learn what allows the resistance, which could be quite valuable. Perhaps this has been done.

As a computer nerd who has worked with very simple genetic algorithms, it is pretty clear to me that new information can be generated by simple mutation. Genetic algorithms are very good at searching for solutions to problems. One only needs a test for goodness of a solution. The more sensitive the test, the more efficient the search. The solutions represent new information in the genome.

Yet life doesn't just use random mutation. The complexities are numerous. For example, genetic information can be moved from one species to another by viruses. Also, since DNA operates at the molecular level, it has access to the molecular machine. It appears to be able to find novel ways to interpret the genetic information available, by altering the interpreter. My computer doesn't do that, because I didn't tell it to.

The complexity of life puts the Evolution debate out of the depth available in the thinking of popular culture. It is desirable that reliance on some authority is poor, so what it takes is an experiment anyone could perform at home with inexpensive equipment, in kit form. Ideally, one could put methane, argon, and some other gunk in a sealed test tube, heat it, cool it, zap it with electricity, etc., and in a few days at most, life would exist, where there was none before. And, this life would be detectable unambiguously. Anyone could do it. And even this would be an incredible stretch for popular culture. Since such an accessible experiment won't be available any time soon, the popular culture is effectively shut out of any such debate. Since it appears that the Intelligent Design debate is squarely pointed at popular culture, and lacks the depth needed for science, that it was never intended to be or become science.

One last point. In the referenced blog, there is some name calling. It starts soon in the blog, but no one took the bait. At post 17, John A. Davidson refers not to Darwinism, but NeoDarwinism as the rival hypothesis. This is quite an insult for someone born in 1928, but probably not for anyone much younger. It refers to Eugenics, with all the racial prejudice and such, and which was used by the Nazis to justify the slaughter of millions in gas chambers. Nice. This was John A. Davidson's second post to this thread, and no one had yet responded to his first. So, it was an unprovoked personal attack. There are actually two attacks here. The second is that he uses the word Darwinism as an ideology. But Darwinism isn't an ideology. Science isn't based on faith, it is based on evidence.

If one edits out the name calling and follows the scientific arguments, there does seem to be real value here. Some of the more subtle Intelligent Design arguments are presented and refuted. One might post such an edit, but one would be in constant fear of misrepresenting someone. For myself, the content is good, but not worth that much effort and risk.

Thursday, December 08, 2005

Bumper

With no crisis at home last night, thoughts drifted over dinner on what should be done to consume the evening. There would be nearly three hours to kill. That's long enough for a movie, and it would be well deserved, given the spate of must-do-it-now tasks that have consumed my waking time for the past week or two. Time for a fun job.

My new car (it's new to me) arrived with the front bumper hanging not quite right with hastily applied bailing wire. It was my bailing wire and tools, though someone else performed the task. The car is in the garage already, and this task doesn't require the engine to cool first, so the task can start right away. One thing that makes the task potentially fun is that though it has been on my TO DO list for some time (really, it would have been more comfortable to do this task in August or September than wait until chilly December), the job was never really very high on the priority list. It isn't a safety issue, it's cosmetic - at least, that was the theory.

Now it should be noted that all the really successful car repair tasks have required that some injury is exacted. It's best if blood is drawn. It helps, too, if some muscles or other are in agony for days. That's my experience. Anecdotal evidence, to be sure, but frequently, if there was no injury, then the repair will require more attention soon. So it is that complete success can be reported due to the chunk of flesh gouged out of my left thumb. Oh, happy day.

So it may be perhaps surprising, but this task was begun with some giddy anticipation. This kind of activity is really quite enjoyable, more or less. A fun job indeed.

The first step was to inspect the damage. After a cursory inspection, it was concluded that to do a better job, the bailing wire needs to be cut. This was done with some reluctance. Cutting it means that the car is no longer drivable, and it has been in use as my commuting vehicle for about a week now. However, my previous car still runs, though it requires considerable repair also. One of the things it needs is a cabin heater blower, so there is little heat, and it is December in Michigan, after all. Still, it is available should this task remain unfinished before bedtime. So with as much courage as could be mustered, the wire was cut. The wedding bells are now audible, my marriage to this job is consummated.

The bumpers on this 2000 Saturn SL are really a fairly thin plastic veneer covering the somewhat more substantial impact point. The idea appears to be to use cheap and easily replaceable parts that will absorb impact energy. The main absorbing material is styrofoam. In an real accident, this bumper will need to be replaced (that is, if the car isn't totaled). So the up side is that replacement won't cost that much. The down side is that any little fender bender may require full replacement. It is not a five mile per hour bumper. This may be news. The US law requiring that bumpers be able to survive five mile per hour impacts was passed with much fanfare. However, it was rescinded without much fanfare. As a guess, most people probably think that a mild impact won't cause any damage to their vehicles.

With the wire cut, the right side of the bumper hung mostly to the floor. There were four plastic screws holding the bumper up in the middle of the car, and they were removed in just a few seconds. The rear bumper was recently fixed, which uses the same sort of fasteners, so my experience accelerated this task. With these removed, the whole right side of the bumper was on the floor. This exposed the styrofoam impact material. Though there is evidence of a front impact, it appears that it was minor, and the impact material is intact and fully functional. It is good to know this. Were it compromised, it would have to be replaced. This is a safety issue, after all. The plastic veneer is not, though it may be aerodynamic and aesthetic. This information might not have made it to me if someone else performed the job.

A further inspection of the veneer shows that it there used to be four plastic screws holding it to the vehicle on the underside. Also on the underside are three attachment points with bolts. There were also four plastic screws holding the right side on, and a metal plate with two bolts securing it to the frame. The plate is gone, along with the bolts. They might have been torn from the vehicle. Yet, the front damage was so slight, that it is possible that the plate was removed manually. Most of the screw holes have been torn through, so even if the screws were available, they wouldn't work. There is also a large tear in the plastic of the right wheel well. The only usable connection is the center underside bolt. Everything else will require a jury rigging somehow.

The center bolt was easy to remove and secure. A jack was used to prop up the bumper while the rest of the task proceeded. The emerging strategy was to cut small holes with a drill, and use bailing wire to tie it to whatever is handy. This worked quite well. A couple of attempts had to be retried, as bailing wire becomes somewhat brittle when twisted too much. Several new techniques were explored, with varying degrees of success. The large tear was repaired with four stitches of wire, for example. These stitches in the inside of the wheel well are the only really visible uses of bailing wire. The job mostly looks professional. Most of the repairs appear to be at least as secure as the original fasteners. The bumper was quite snugly attached on the right side. The missing metal plate wasn't replaced, so this was the most worrisome part. All in all, a good job, with a total cost of $0.00.

The total task exceeded the available time by an hour. That just means that bed time was pushed back by that much. This was worth it, as the job was finished, leading to some satisfaction. Besides, it would be available for the morning commute, and the thick frost on the other car would not have to be scraped off. Yet, there still was a lingering negative feeling, which took a few minutes to identify. And my wife is entirely to be blamed for it. On multiple occasions after finishing such a task, she would complain with a litany of other tasks that went undone. It was as if my time was wasted on this task instead of others. Demoralizing doesn't cover this attitude. Even though this behavior was stopped entirely over a year ago, the long term damage is done. It was the gift that keeps on giving. Here's the real damage. Rather than perform this task which gives me real pleasure, and which turns out to be vital to the running of the house, i'm often so depressed, i'd rather do nothing of value so that i'm not wasting my time, not doing some other task. Feh. There are many fun things one could be doing at any given time in life, and you are missing most of them.

Wednesday, December 07, 2005

Chicken and Egg

Which came first, the Chicken, or the egg? When i a little kid, and was asked that question, i answered fairly quickly - the egg. My reasoning then was that chickens are likely a relatively modern species, and that there were likely older species that you wouldn't call chickens that laid eggs.

Of late, a new Archaeopterix fossil, with a well preserved foot, strongly suggests that birds are descendant from dinosaurs, with much less possibility that they predate dinosaurs - perhaps being completely missed in the fossil record. The foot isn't a modern bird style perching foot with a hind toe that points backwards. It looks more like a dinosaur foot, which leaves something for birds to evolve on their own. In any case, if dinosaurs came first, then since they laid eggs, the egg came before all birds, not just chickens. That's what i suspected when i was a little kid. Now i've seen evidence that my hypothesis was correct.

Note - i prefer to say that the evidence suggests rather than i believe, if for no other reason than that if better evidence becomes available, i can say that i wasn't wrong given the evidence available, its just that the new evidence is better, and, no, i haven't changed my mind, i've always tried to use the best evidence. Also, i'd like to think that i'm trying to learn science, rather than just religion. Science isn't a matter of faith. In my opinion, scientists don't believe in Evolution, the evidence suggests (demands) that this theory is correct. While there could be a theory that fits the evidence better, none has been suggested. The nay sayers might say, Evolution is just a theory, however, in science, a theory has lines of reasoning, but also a body of evidence that supports it. A hypothesis is the thing that just has the lines of reasoning. So, the nay sayers are suggesting that Evolution is just a hypothesis, which is silly. Evolution is more like gravity. It doesn't speak, but commands attention.

Tuesday, December 06, 2005

With Malice, toward none

I mentioned that I'd been doing some reading on Intelligent Design over lunch. I said that I hate to refer to these people as ID'ers, because they aren't the designer(s) in question.

The reaction was "nor are they intelligent". Also, "Never attribute to malice what can be attributed to stupidity."

My reply was that at least one ID'er was obviously very careful, and managed to use, on average, more than one specious argument (pun intended without apology) per sentence. It's hard to even know where to begin to argue with it. With someone that smart, concluding malice is unavoidable.

To which, the guys said, "Well, then, they're dammed to Hell."

Thursday, December 01, 2005

End User License Agreement

I'd heard this some time ago as a Swedish joke. Its really a Stupid joke, rather than a slam on any nationality. There is no attempt here to make it generic. Generic jokes end up like this:

Ethnic #1 goes to her Rabii...

On with the joke.

Ollie visits Sven, who is busy in the back yard sawing wood with a cross cut saw. Ollie says to Sven, "Sven, that looks like hard work. Why don't you get a chain saw?" So Sven goes to Sears & Robuckson and buys a chain saw. Later Sven is talking to Ollie. "That chain saw is no good. Its even more work than the cross cut saw. I'm going to return it." Ollie is amazed and says, "Sven, you've got to show me this chain saw. They're wonderful. There's no way it won't save you work." So the next day Ollie visits Sven and Sven shows Ollie the unit. Ollie says, "It looks like a good one to me". He picks it up and pulls the chord. The chain saw roars to life. Sven says, "What's that noise?"

When people purchase new stuff, a computer, an iPod, a child's toy, they often come to me with questions on how to get some feature to work. These are people who see me pick up some item and make it do something they've never seen it do. They figure i'm some sort of technology genius, and figure i know everything. By the time they ask questions, the documentation, which has never been opened, is gone - misplaced or even thrown out. Sometimes the item has been partially disassembled, with parts missing. They expect me to fix it. Its really frustrating, as its often something i've never seen before.

When i ask for the documentation i sometimes get puzzled looks. "Nobody reads that stuff". And i reply that i even read the Warranty and License Agreements.

I'm not a lawyer, so maybe i don't get all the nuances out of the License Agreements. Most of them say that you must agree that the seller is absolved of any liability stemming from your use of the product, within the limits established by local law (which may not absolve them of any liability at all). Typically these License Agreements only pertain to software aspects of the product. Software is treated differently from other works of engineering. Why is that?

One reason is that software makers demand it. Software is hideously complicated, and errors are difficult to detect and impossible to avoid. The software that guides the Space Shuttle into space is thought to have around 50 errors. One might expect that if such a thing is known that every effort would be expended to find these bugs and fix them, so that we don't lose a multi billion dollar orbiter and seven astronauts. The fact is that the bug count is an estimate. There is no direct evidence for any of these bugs. And every expense has been spent looking for them - something like $10,000,000,000. Even Microsoft does not have enough money to eradicate every bug from Windows. If Windows had to be bug free in order to go on the market, it wouldn't go on the market. Hence the demand to be exempt from liability.

Why is software so hard to make error free? Well, for one thing, its more complicated than any other engineering task. A 747 has something like 10,000,000 fasteners. But most of these look like each other, with only dozens of different kinds. In contrast, software consisting of a million lines of code would have a dozen routines acting like glue, which would be called many times throughout the rest of the code. Very few sections of the code would look like other sections. So there are tens of thousands of modules, all of which may interact in odd ways with every other module. The 747 is looking pretty simple by comparison.

Software entities are more complex for their size than perhaps any other human construct because no two parts are alike. If they are, we make the two similar parts into a subroutine -- open or closed. In this respect, software systems differ profoundly from computers, buildings, or automobiles, where repeated elements abound. - Fred Brooks


The great Fred Brooks is one of the most ignored luminaries in the practice of computer science. Very sad, really.

So, how do you protect yourself from buggy software? Easy. Buy simpler devices. The number of bugs appears to grow with the size (complexity) of the code. Simpler is better.