Today's technology may be nailing down more accurate dating, but human experience out in the field is still you're best place to start in an archaeological dig. While the two should compliment each other, the people who rely on machines to do all the work for them don't really understand what it means to be an archaeologist.
My field work has suggested to me that "what it means to be an archaeologist" is basically sitting around and spouting lots of complex scenerios from rather trivial amounts of data. Whoever pushes the grandest vision gets the biggest funding and that's what determines our view of reality. I thought about going into the field at one time but was quickly discouraged. While it's an interesting (and useful) mental exercise to sit around and think up these scenerios, I think the amount of real hard knowledge that we derive from them is essentially meaningless. Fact is, pretty much all of what we know comes from written records. The rest, we owe mostly to the harder sciences (carbon dating, genetic analysis, etc). And the speculation, well, that's not so easy to test...
I was using the Wild West as a metaphor. A very large part of the American personality is the "the rules don't apply to me" or the "who's going to stop me" attitude. An American will do what he can until he meets resistance and even then will try to find a way around it. That's what makes the US an incredibly creative place where everything is possible. But here in Switzerland people live according to a social morality. They don't just do what they want but they consider whether it's "right" or not (where "right" is usually defined as the good of society). As a result, life moves slower here but it's a healthier environment. I think this difference came out of the lack of structure in the US in the early days and, hence, my term "Wild West". As for everyone wanting to be JD Rockefeller and not Jesse James... it would be hard to objectively define the difference. 8-)
Devon
Just please educate me. What is so wrong about the card?
What's wrong isn't the card, exactly. It's more a mismatch between the culture and the card concept.
I don't have any trouble registering with my Gemeinde here in Switzerland because I know that this information is respected and secured by the government and the people. Swiss people don't think "hey, how can I exploit this for money or power?" It's not part of the culture. And, as such, I feel that I have essentially nothing to worry about.
In America, however, the first thing that pops into anyones head is "hey, how can I exploit this for money or power?" It's the American mindset that grew out of the Wild West and is still strong. There is no way in hell I'd want to register with the my local police department in the US. Because I know what would come later.
The card itself is a tool. It can be used for good or bad. The culture determines which.
Is it just me or have others noticed that a large fraction of Slashdot stories come from CNN these days? I read Slashdot for things that I can't find in popular media. Otherwise, what's the point?
not to be confused with the biometric technology of retinal scanning, which detects the distinctive pattern of blood vessels on the retina, a pattern that can be analyzed to identify someone, similar to the way fingerprints are used.
Why exactly should they be together?
Devon
A friend of mine was using Napster to download MP3 versions of all his old records. He already paid for the use of this music when he purchased the original. Preventing copying of music via encryption would force him into a cycle of paying for the same thing as often as the industry wants to create a new format. The industry would also be encouraged to produce poor quality products so that they would need to be replaced often.
Beings with ten digits don't normally get excited about 33rd anniversaries
Which doesn't make much sense, actually. Ten digits allows us to represent eleven numbers (0 through 10) so we should have a base eleven counting system. But for some reason, the ancients decided to roll over at ten even though our fingers don't. And, no, it's not because they didn't understand the concept of zero. This is an issue of when to roll over, not what to use as a place holder. The way we write "10" also requires zero as a place holder.
Now, about the above point of iterative/functional making more sense, naturally. Firstly, think about how you tend to solve problems: come up with an idea, walk through the steps to get it to work. Note that there is nothing about objects in there. A pseudo-code algorithm translates a lot easier to a C program than it does a Java program.
I see what you're getting at but I think it might be a problem with teaching style. To paraphrase, you're saying that you can't take a procedural way of thinking and map it into an OO world. With this I would agree.
But instead of teaching the students to think in terms of procedures from the start, teach them to think in terms of objects. Teach them that building a program is like building a clock. You take parts and put them together in a certain way and the "procedure" falls out of that structure just as it does within a clock. Now we can rewrite what you said as:
Now, about the above point of OO making more sense, naturally. Firstly, think about how you tend to build machines: come up with an structure, plug in the parts to get it to work. Note that there is nothing about procedures in there. An OO model translates a lot easier to a Java program than it does a C program.
Then teach them that in software you can't usually order the parts from someone else and you need to build those, too. Work your way down into the tiny procedures needed inside your objects methods. This puts the emphasis in the right place - on the overall model and structure. A top down approach makes a whole lot more sense for teaching OO.
They ignored Tesla when he said that alternating current was the way to go, and gave edison all the recognition for bringing electricity to the public, even though he tried and failed to do it with DC, and Tesla's proposed AC system ended up being the key. This is another case where the inventer of a cool technology is for the most part ignored.
While I agree that Tesla was screwed over in a number of ways, this wasn't one of them. Tesla didn't invent alternating current but rather a number of inventions that made AC usable. Because of these inventions he (under Westinghouse) won the rights to build the electrical infrastructure at the Chicago Worlds Fair in 1893. Shortly thereafter, he was won the rights to outfit Niagara Falls for the generation of AC power (statue of him stands there today). Tesla's name was a household word at that time and he was given all the credit in the world quickly overtaking Edison (who was first, actually) in the war to bring electricity to the people.
Rebel Code, by British author Glyn Moody is one of the first serious histories of this movement.... Moody begins the book at the peak of Microsoft's rule, with the primal beginnings of Linux at the hands of Linus Torvalds, then a college student in Finland.
A serious history about the open source movement cannot possibly begin with Microsoft and/or Linux. My students and I had already donated half a dozen GNU ports to the well-established FSF before then. Berkeley had already released major parts of BSD and were in the process of rewriting parts still tied up by AT&T licensing before then. Hundreds, if not thousands, of software packages had been developed and posted to comp.source newsgroups before then. Get real.
The arrogance of Linux brats never ceases to amaze me.
Chaos theory is the study of systems which have very simple equations of motion, but which have an extremely compilcated behaviour.
Chaos theory is the study of systems with a behavior such that an arbitrarily small change in input parameters results in an output that is not at all close to the previous result. For example, a weather forecastor which, when given an input of 5 degrees, predicts a thunderstorm yet when given an input of 5.001 degrees predicts sunshine. In fact, Chaos theory was born from just such an example.
Chaos is not about "simple equations of motion" although a number of these types of equations are used as examples.
First, serious people in the community generally shun the word "spelunker" which was coined by journalists for journalists. The preferred term is "caver". When someone describes themselves as a spelunker, I usually proceed with caution. It's often noted that spelunkers are rescued by cavers.
As for the access to such precious beauties... there is plenty of precedent in this area. Many caves have a scientific use beyond "hey, this is cool". For example, there is a cave in Romania (Movile) that contains an entire ecosystem adapted to a toxic environment. Or the microscopic flora in Lechuguilla which is being studied for its stronger than usual cancer fighting abilities. There are strong regulations in both of these places even against visitation by serious spelunkers and cavers alike. It's also not unheard of to close a cave for geologically scientific reasons.
Then there's places like Lascaux, the most famous of the prehistorically painted caves, which have been painstakingly copied so that people can visit the replicas without polluting the original. While thousands may visit the replica each day, the original now has a maximum visitation of five people per day by appointment only.
There are plenty of alternative ways of thinking about caves. In the case of free software, you've got essentially infinite resources. But with caves, there is often a rarity and fragility that requires something more than a simple "everyone gets equal access" approach.
BTW, if you're looking for some really interesting cave reading, check out "Dark Life" by Michael Ray Taylor (author of the crystal article). He talks about the connection between Martian nanobacteria and the research done in caves today. A good way to spend a day.
I wouldn't exactly say that zero-energy computation isn't anything new. It is true that arbitrarily-small-amounts-of-energy-computation has been possible but only with the trade-off of time. The less energy you use the more time is required. Zero-energy would require infinite amount of time which isn't too useful.
The specific equation that governs this states that the minimal energy required for a computation is k*T*log(r) where k is the Boltzmann constant, T is temperature, and r is the rate of computation (the ratio of forward computations to backward computations in a reversible gate). Note that the only time the energy is zero is when the rate of computation is one. Since this is a ratio, it essentially means that there is no more forward computation than backward which means you'll never get an answer to your problem.
For more information on this topic, I highly recommend "Feynman Lectures on Computation" by (can you guess?) Richard Feynman. In fact, I highly recommend this book for anyone interested in the science of computation.
What possible reason can anyone give for not dumping LinuxPPC? With MacOS X I'll have a nice graphical environment, I'll be able to buy third party commercial packages, and, with the addition of X windows, I'll have all the Linux tools that I need. When a new hardware device comes out (FireWire, USB, DVD - all adopted early by Apple) I'll have drivers for them immediately. Sure it'll cost a bit more than a LinuxPPC disc set but that isn't significant. The only reason for running LinuxPPC would be to appear cool on slashdot.
One important thing to keep in mind is that times change. You can't just look at the current situation when making your decision. In today's world there is such a desparate need for technical people that just spelling "Java" will get you a job. Don't laugh. I've seen it. And I am talking about very high paying jobs.
But that's the world right now - a very short time after the peak of dotcomness. Imagine three years from now when we're in a heavy recession. Money is tight and people aren't so willing to start new projects and take on new people. It's that situation that you should be thinking about.
CS degrees are harder because you're learning a lot more. This will make you a lot more marketable when push comes to shove. In today's world that probably isn't a huge issue and I know that it's hard to imagine the economy any other way after recent years. But bad times will come again. As sure as the setting sun. Just something to keep in mind.
For a given problem there will be a set (or sets) of models (given some restraints) which will give an optimal solution... are you saying its provable there is no faster method of determining those sets than with the random generation method?
I don't think that was the point that "junkmaster" was trying to make. But there is a certain advantage to using randomly generated sets.
The mathematics of sd shows that by increasing the strengh of the weak models that are "combined", one requires fewer of these weak models to get the same quality of recognition. In fact, there is even an equation that puts limits on this (which I believe is based on Chebyshev's inequality, but I don't remember exactly) and the most commonly used implementation of sd does allow you to set a threshold for the quality of weak models you want chosen from those that are randomly generated.
However, remember that the concept of uniformity is the linchpins in sd. All the models that you choose to "combine" must be uniform, as defined by the theory, with respect to each other and to the problem space. Randomly chosen models tend to make it easier to accomplish this because the fact that they are random already gives them a closer to uniform coverage of the feature space. They just need to be tuned to get true uniformity.
There is nothing theoretically wrong with using stronger weak models to get to your solution faster. But by putting "intelligence" into the process of making the weak models, it would be much, much harder to ensure that they were mathematically uniform. I said it in my original post but I should stress the point that the uniformity concept is the reason that sd beats out all the other methods for "combining" weaker models in standardized benchmarks. The importance of obtaining it cannot be underestimated.
As already mentioned, the concept of "combining" the results of pattern recognition models is not new and there have been various techniques for doing this. Most of these techniques have been ad hoc with very little rigorous mathematical foundation.
One notable exception is in the research of Stochastic Discrimination (sd). This technique was originally developed through mathematics rather than experimentation as is the case with NNs. In other words, rather than the "let's see what happens if" development of NNs, sd's approach is "the equations say this should happen". Because of this, it is very rigorously defined and the hows and whys are clearly understood.
Sd also "combines" weak models but in a way that, to the best of my knowledge, no one else has done before. The basics are:
Incredibly weak models are generated to solve the given problem.
Hundreds of thousands of these models are combined.
These weak models must be uniform with respect to each other and to the problem space.
For example, rather than combining a few very specific models as described in this article (one for furry animals, one for domestic, etc), sd would randomly generate hundreds of thousands of weak models to solve the problem. Each of these models would look at a different set of features but there would be so many combinations that you wouldn't be able to name them. For example, maybe one model would learn to distinguish based on the length of tail, the color of snout, and diameter of the third toenail. This model obviously can't be named. The set of features it looks at are too odd. But if we note that it has some trivially weak ability to tell the difference between a dog and a cat then we accept it. For the problem of dogs vs cat, we may only require that any given model be 50.1% accurate on our training set. When we "combine" all these weak models, a strong solution emerges. Why this happens has its roots in the Central Limit Theorem.
But before we "combine" them, we have to see that this weak model is uniform with respect to the other weak models. This is a term defined in the sd theory. Basically, what it means is that the weak models need to be evenly selected throughout the set of all possible weak models. In other words, there is no oversampling or bias. (Actually, this isn't quite right but goes in the right direction. Read one of the papers if you're interested.) The concept of uniformity is probably the most interesting part of sd and it is the primary concept that all the other "combination" techniques miss. In this article, for example, how do we know that there isn't a connection to being furry and being small? If there is a statistical dependency, then the vote won't be fair and results will be weaker.
Anyway, that's a real crash course in sd basics. So how does this algorithm perform? On the standard benchmarks (Irvine, for example) it handily outperforms anything out there. Right out of the box and without tuning. For more information see the web site or send me email.
Re:Use encryption needlessly, constantly! [MUCH MO
on
The Encryption Wars
·
· Score: 2
Everyone knows that encryption is a good thing. Even my mother knows because the media has been hyping the lack of security on the net for years now. For example, while not at all a savy user, she has actively asked me to show her how to know when a connection is secure enough to send her credit card number. All this from just listening to the media.
The problem isn't awareness. The problem is the user interface. My mother knows how to make her postal mail secure. She simply opens an envelope, puts in the letter, and licks it shut. The person on the other end also knows how to get at the letter. She doesn't need to explain to them how to open the envelope. This isn't true for email. I can't even explain to her how to send encrypted mail because it's just too difficult. And there's absolutely no way she could explain it to her friends. So even if she could send it, it couldn't be read by the recipient.
Any change has to happen transparently. She has to send the letter to joe@foo.com, the public key for joe@foo.com needs to be automatically looked up in one or more global databases, and it needs to be encrypted and sent. All without her even noticing. If joe@foo.com hasn't registered a public key then it should tell her the mail is insecure and to tell the other user to register a public key. The other user should also get a warning. Probably at the start of the email.
Why this functionality isn't already built into every single mail reader in existence, I just don't understand. Some people will raise the issue of needing to decentralize the storage of the keys, etc, but isn't that an incredibly minor concern considering the state of things now? People make secure connections all the time on the net. They do this because SSL is built into every browser and encryption happens transparently and automatically. Until we have the same for email, we won't have secure email.
Slashdot Mirror
My field work has suggested to me that "what it means to be an archaeologist" is basically sitting around and spouting lots of complex scenerios from rather trivial amounts of data. Whoever pushes the grandest vision gets the biggest funding and that's what determines our view of reality. I thought about going into the field at one time but was quickly discouraged. While it's an interesting (and useful) mental exercise to sit around and think up these scenerios, I think the amount of real hard knowledge that we derive from them is essentially meaningless. Fact is, pretty much all of what we know comes from written records. The rest, we owe mostly to the harder sciences (carbon dating, genetic analysis, etc). And the speculation, well, that's not so easy to test...
Devon
I was using the Wild West as a metaphor. A very large part of the American personality is the "the rules don't apply to me" or the "who's going to stop me" attitude. An American will do what he can until he meets resistance and even then will try to find a way around it. That's what makes the US an incredibly creative place where everything is possible. But here in Switzerland people live according to a social morality. They don't just do what they want but they consider whether it's "right" or not (where "right" is usually defined as the good of society). As a result, life moves slower here but it's a healthier environment. I think this difference came out of the lack of structure in the US in the early days and, hence, my term "Wild West". As for everyone wanting to be JD Rockefeller and not Jesse James... it would be hard to objectively define the difference. 8-) Devon
What's wrong isn't the card, exactly. It's more a mismatch between the culture and the card concept.
I don't have any trouble registering with my Gemeinde here in Switzerland because I know that this information is respected and secured by the government and the people. Swiss people don't think "hey, how can I exploit this for money or power?" It's not part of the culture. And, as such, I feel that I have essentially nothing to worry about.
In America, however, the first thing that pops into anyones head is "hey, how can I exploit this for money or power?" It's the American mindset that grew out of the Wild West and is still strong. There is no way in hell I'd want to register with the my local police department in the US. Because I know what would come later.
The card itself is a tool. It can be used for good or bad. The culture determines which.
Devon
Devon
And from the posted article:
not to be confused with the biometric technology of retinal scanning, which detects the distinctive pattern of blood vessels on the retina, a pattern that can be analyzed to identify someone, similar to the way fingerprints are used. Why exactly should they be together? Devon
Devon
Which doesn't make much sense, actually. Ten digits allows us to represent eleven numbers (0 through 10) so we should have a base eleven counting system. But for some reason, the ancients decided to roll over at ten even though our fingers don't. And, no, it's not because they didn't understand the concept of zero. This is an issue of when to roll over, not what to use as a place holder. The way we write "10" also requires zero as a place holder.
Devon
I see what you're getting at but I think it might be a problem with teaching style. To paraphrase, you're saying that you can't take a procedural way of thinking and map it into an OO world. With this I would agree.
But instead of teaching the students to think in terms of procedures from the start, teach them to think in terms of objects. Teach them that building a program is like building a clock. You take parts and put them together in a certain way and the "procedure" falls out of that structure just as it does within a clock. Now we can rewrite what you said as:
Now, about the above point of OO making more sense, naturally. Firstly, think about how you tend to build machines: come up with an structure, plug in the parts to get it to work. Note that there is nothing about procedures in there. An OO model translates a lot easier to a Java program than it does a C program.
Then teach them that in software you can't usually order the parts from someone else and you need to build those, too. Work your way down into the tiny procedures needed inside your objects methods. This puts the emphasis in the right place - on the overall model and structure. A top down approach makes a whole lot more sense for teaching OO.
Devon
While I agree that Tesla was screwed over in a number of ways, this wasn't one of them. Tesla didn't invent alternating current but rather a number of inventions that made AC usable. Because of these inventions he (under Westinghouse) won the rights to build the electrical infrastructure at the Chicago Worlds Fair in 1893. Shortly thereafter, he was won the rights to outfit Niagara Falls for the generation of AC power (statue of him stands there today). Tesla's name was a household word at that time and he was given all the credit in the world quickly overtaking Edison (who was first, actually) in the war to bring electricity to the people.
Devon
A serious history about the open source movement cannot possibly begin with Microsoft and/or Linux. My students and I had already donated half a dozen GNU ports to the well-established FSF before then. Berkeley had already released major parts of BSD and were in the process of rewriting parts still tied up by AT&T licensing before then. Hundreds, if not thousands, of software packages had been developed and posted to comp.source newsgroups before then. Get real.
The arrogance of Linux brats never ceases to amaze me.
Devon
Chaos theory is the study of systems with a behavior such that an arbitrarily small change in input parameters results in an output that is not at all close to the previous result. For example, a weather forecastor which, when given an input of 5 degrees, predicts a thunderstorm yet when given an input of 5.001 degrees predicts sunshine. In fact, Chaos theory was born from just such an example.
Chaos is not about "simple equations of motion" although a number of these types of equations are used as examples.
Devon
As for the access to such precious beauties... there is plenty of precedent in this area. Many caves have a scientific use beyond "hey, this is cool". For example, there is a cave in Romania (Movile) that contains an entire ecosystem adapted to a toxic environment. Or the microscopic flora in Lechuguilla which is being studied for its stronger than usual cancer fighting abilities. There are strong regulations in both of these places even against visitation by serious spelunkers and cavers alike. It's also not unheard of to close a cave for geologically scientific reasons.
Then there's places like Lascaux, the most famous of the prehistorically painted caves, which have been painstakingly copied so that people can visit the replicas without polluting the original. While thousands may visit the replica each day, the original now has a maximum visitation of five people per day by appointment only.
There are plenty of alternative ways of thinking about caves. In the case of free software, you've got essentially infinite resources. But with caves, there is often a rarity and fragility that requires something more than a simple "everyone gets equal access" approach.
BTW, if you're looking for some really interesting cave reading, check out "Dark Life" by Michael Ray Taylor (author of the crystal article). He talks about the connection between Martian nanobacteria and the research done in caves today. A good way to spend a day.
Devon
The specific equation that governs this states that the minimal energy required for a computation is k*T*log(r) where k is the Boltzmann constant, T is temperature, and r is the rate of computation (the ratio of forward computations to backward computations in a reversible gate). Note that the only time the energy is zero is when the rate of computation is one. Since this is a ratio, it essentially means that there is no more forward computation than backward which means you'll never get an answer to your problem.
For more information on this topic, I highly recommend "Feynman Lectures on Computation" by (can you guess?) Richard Feynman. In fact, I highly recommend this book for anyone interested in the science of computation.
Devon
Devon
But that's the world right now - a very short time after the peak of dotcomness. Imagine three years from now when we're in a heavy recession. Money is tight and people aren't so willing to start new projects and take on new people. It's that situation that you should be thinking about.
CS degrees are harder because you're learning a lot more. This will make you a lot more marketable when push comes to shove. In today's world that probably isn't a huge issue and I know that it's hard to imagine the economy any other way after recent years. But bad times will come again. As sure as the setting sun. Just something to keep in mind.
Devon
I don't think that was the point that "junkmaster" was trying to make. But there is a certain advantage to using randomly generated sets.
The mathematics of sd shows that by increasing the strengh of the weak models that are "combined", one requires fewer of these weak models to get the same quality of recognition. In fact, there is even an equation that puts limits on this (which I believe is based on Chebyshev's inequality, but I don't remember exactly) and the most commonly used implementation of sd does allow you to set a threshold for the quality of weak models you want chosen from those that are randomly generated.
However, remember that the concept of uniformity is the linchpins in sd. All the models that you choose to "combine" must be uniform, as defined by the theory, with respect to each other and to the problem space. Randomly chosen models tend to make it easier to accomplish this because the fact that they are random already gives them a closer to uniform coverage of the feature space. They just need to be tuned to get true uniformity.
There is nothing theoretically wrong with using stronger weak models to get to your solution faster. But by putting "intelligence" into the process of making the weak models, it would be much, much harder to ensure that they were mathematically uniform. I said it in my original post but I should stress the point that the uniformity concept is the reason that sd beats out all the other methods for "combining" weaker models in standardized benchmarks. The importance of obtaining it cannot be underestimated.
One notable exception is in the research of Stochastic Discrimination (sd). This technique was originally developed through mathematics rather than experimentation as is the case with NNs. In other words, rather than the "let's see what happens if" development of NNs, sd's approach is "the equations say this should happen". Because of this, it is very rigorously defined and the hows and whys are clearly understood.
Sd also "combines" weak models but in a way that, to the best of my knowledge, no one else has done before. The basics are:
- Incredibly weak models are generated to solve the given problem.
- Hundreds of thousands of these models are combined.
- These weak models must be uniform with respect to each other and to the problem space.
For example, rather than combining a few very specific models as described in this article (one for furry animals, one for domestic, etc), sd would randomly generate hundreds of thousands of weak models to solve the problem. Each of these models would look at a different set of features but there would be so many combinations that you wouldn't be able to name them. For example, maybe one model would learn to distinguish based on the length of tail, the color of snout, and diameter of the third toenail. This model obviously can't be named. The set of features it looks at are too odd. But if we note that it has some trivially weak ability to tell the difference between a dog and a cat then we accept it. For the problem of dogs vs cat, we may only require that any given model be 50.1% accurate on our training set. When we "combine" all these weak models, a strong solution emerges. Why this happens has its roots in the Central Limit Theorem.But before we "combine" them, we have to see that this weak model is uniform with respect to the other weak models. This is a term defined in the sd theory. Basically, what it means is that the weak models need to be evenly selected throughout the set of all possible weak models. In other words, there is no oversampling or bias. (Actually, this isn't quite right but goes in the right direction. Read one of the papers if you're interested.) The concept of uniformity is probably the most interesting part of sd and it is the primary concept that all the other "combination" techniques miss. In this article, for example, how do we know that there isn't a connection to being furry and being small? If there is a statistical dependency, then the vote won't be fair and results will be weaker.
Anyway, that's a real crash course in sd basics. So how does this algorithm perform? On the standard benchmarks (Irvine, for example) it handily outperforms anything out there. Right out of the box and without tuning. For more information see the web site or send me email.
The problem isn't awareness. The problem is the user interface. My mother knows how to make her postal mail secure. She simply opens an envelope, puts in the letter, and licks it shut. The person on the other end also knows how to get at the letter. She doesn't need to explain to them how to open the envelope. This isn't true for email. I can't even explain to her how to send encrypted mail because it's just too difficult. And there's absolutely no way she could explain it to her friends. So even if she could send it, it couldn't be read by the recipient.
Any change has to happen transparently. She has to send the letter to joe@foo.com, the public key for joe@foo.com needs to be automatically looked up in one or more global databases, and it needs to be encrypted and sent. All without her even noticing. If joe@foo.com hasn't registered a public key then it should tell her the mail is insecure and to tell the other user to register a public key. The other user should also get a warning. Probably at the start of the email.
Why this functionality isn't already built into every single mail reader in existence, I just don't understand. Some people will raise the issue of needing to decentralize the storage of the keys, etc, but isn't that an incredibly minor concern considering the state of things now? People make secure connections all the time on the net. They do this because SSL is built into every browser and encryption happens transparently and automatically. Until we have the same for email, we won't have secure email.
Devon