I guess I should have said, "The region of the Laffer Curve where increased econonomic activity will compensate overall tax revenue lost by reducing the tax rate." Reagan clearly wasn't on that part of the Laffer Curve, either. And perhaps that wasn't really the problem, it's that neither Reagan nor Bush II cut taxes while keeping spending under control.
Is it so they can drop support, or is it so they can drop the one reliable method of moving on-machine files between Windows and Linux?
FAT compatilibity is stunningly important to Linux. Assuming much Linux migration begins as dual-boot, FAT stitches the two sides of the box together. FAT is how we read/write those silly digital cameras across platforms. FAT is how we read/write those silly memory keys across platforms. FAT is how we read/write diminutive floppies across platforms.
I was just futzing with this last night on my files that came from a Sony digital camera. If the patents truly are related to long filenames, looks like Sony might be covered, unless they support long directory names (Entering LFNs them using the UI of a digital camera would be a royal pain.) or unless the name gets one digit longer on my millionth picture, assuming the camera lasts that long and I insist on the ability to store every picture I take with it in one directory.
This is a spike. It'll last until the deficit catches up to us, presumably after November, 2004. I'm just trying to figure out when to make my stock moves. The economy isn't doing SO well that it'll put us onto the Laffer Curve.
That was the intent, and I suspect to some extent it still works that way. But I suspect that there's far more happening of a different sort with patents - and that's the creation of a club. Remember the phrase, "stand on the shoulders of giants?" Well, the shoulders are patented. If you want to stand on those shoulders and reach higher, you have to let the giant reach that high, too.
In essence, patents have created a club, and while you can still get a patent and make money, you probably can't disrupt an existing technology, because you need to license existing technology to make your patent work, and the most likely license term is to cross-license your technology back to the would-be disruptees. They can either take advantage of the technology, or you'll find that the license prevents you from disrupting their business - unless you're excessively lucky.
I recently heard about a guy with some sort of chemical/drug/food (forget which) patent that's running out. NONE of the industry has agreed to license it, they're just waiting for it to expire. In the meantime he's losing all of his development and attempted marketing money. Maybe he was asking absurd terms, maybe he deserved them, but the industry felt we could get along without the new product, the guy couldn't commercialize without more money than he had, so they could afford to wait.
Come to think of it, I've got a friend in the very same situation.
What about a 'transfer station' in a highly elliptical orbit around both Earth and Moon? Agreed that a rendezvous with the transfer station would be every bit as energetic as getting to the Moon, but it could be done with something minimal like a Soyuz capsule, in terms of crew space. Once docked, the crew space of the transfer station could be much more spacious, maybe even with some equipment for science. Consumables would have been carried aboard the rendezvous capsule, with the crew. Upon return from the Moon, the lower mass capsule could again be used to get to more Earthly speeds.
Of course the same orbital mismatch would occur at the Moon, perhaps solved in a similar fashion. The transfer station would keep some sort of reusable LEM docked, except when being used on the Moon. That LEM would have the additional requirement of reaching lunar escape velocity in order to rendezvous with the transfer station.
Take the idea a little further - space stations in orbit around both Earth and Moon, and a transfer station orbiting both. Get from Earth to the Earth station, go from there to the transfer station, from transfer station to the Moon station, from there to the Moon. The essence is to get things down to minimal, efficient, single-purpose vehicles, and accelerate as little mass as possible at any given step.
Sure there have been security breaches, security breaches will always be with us. Anyone who can't accept that is probably a prime customer for Palladium, and deserves what they get.
The real issue here is that Debian and Gentoo were both forthcoming about the breaches. They both did the Right Thing. Not only that, but they've both collected forensics, and if not identifying culprits, are at least contributing to improving the security of the Linux community.
This is Real Security, as opposed to hiding the facts, and hoping nobody ever finds out.
There are some who theorize (Sorry, no reference, I read this years (decades) ago in dead tree form.) that life began bootstrapped on clays.
Some clays at the microscopic scale carry electrostatic charges, and at that level can be somewhat self-assembling through sedimentation. So there is a micro-structure there. Now take the water that that microstructured clay is immersed in (We are talking about sedimentation, after all.) and put some organic impurities in it. Those impurities will tend to self-align on the clay face, drive by diffusion, micro-scale surface topology, and electrostatic charge. At some point you may have a sufficiently advanced scum on the clay surface that if a chance current breaks it off, it can 'do something' on it's on - a sort of proto-life.
This scenario is a case of a self-assembled unit that assembles something else.
On the serious side, before getting to the fantasy below...
There is a fuzzy dividing line between the macro and microscopic worlds. Various experiments have been able to move that line, AFAIK always upwards, but the line is still there. At some point, materials quit acting classically and start acting quantumly. Molecular assemblers are clearly below that line.
It's really hard to 'work your way down' to the molecular assembler level, because you're either classical or quantum. Even the experiments that elevate quantum effects into normally classical space are highly specific, and could not likely be extended to an assembler.
The chicken-and-egg problem mentioned elsewhere has a solution. Some 'proto-chicken' laid the first 'true-chicken' egg that hatched and became the first chicken. Rather obvious with a little thought.
The molecular assembler isn't so easy, and building the bridge is tough. I see two ways, either building really simple assemblers in bulk, in 'test tubes' somewhat as we build lipospheres, today, and using them to build more complex assemblers. Or building simple, gross molecular assemblers out of the smallest classical assemblers (STM on steriods?) we can make.
In either case, it's a bootstrap process, and by the time we've bootstrapped we may find that we're not exactly where we expected to be. Not necessarily better or worse, just different. Pre-bootstrap expectations are likely to be wrong.
A pet fantasy from my high school days (early 70's) was to build a smaller Waldo-style robot. (mimic my movements, at another place/scale) Use that Waldo to build a yet smaller one, and so on.
The final goal of this progression was to go walking down the grooves of my favorite vinyl LPs, carrying a micro-trowel and a micro-bucket of vinyl patch, and fix the really annoying clicks and pops by hand.
This was never a serious fantasy, but a fun thought. Not only did I not figure the needed size of the Waldo, beyond the 1-mil (25.4 uM) scale, I didn't figure the length of the virtual hike it would take for one LP.
Not only were vinyl LPs obsolete before nanotech (It didn't have that name, then.) truly approached, my old 'temporary, until I can afford better' AR-XA turntable outlasted the format.
Not to stretch a metaphor, but what does it mean that you can't stay in the eye of the storm by standing still or staying where you are? To stay in the nice weather you have to move with the storm - rather exactly.
Won't argue with you on this one. Assuming the damage is real, selling it gives Party B immediate revenue, allowing it to focus on its core business.
I guess there can be different ranges of 'ethical-ness' for Party C, and it might be questionable whether I'd invite ANY of them over for dinner, though I'd be more likely to invite your first group of Party C than your second group.
Of course developers should Listen to all feedback.
How much of that feedback they should Act on is an entirely different matter. It's probably more important to get some sort of demographic feel for their customer base, to know if it's some sort of lunatic fringe making the most noise, or if complaints are broad-based.
Perhaps the hardest challenge would be developing a game that can appeal both to a broad base, and to hard-core lunatic fans.
One can examine issues and take sides based on them, rather than on the players. Try a search on "aard" "microsoft" and "drdos" for more information. Years back, Dr. Dobbs ran a particularly condemning article on the AARD code. From what I could see, wasn't much 'alleged' about it, other than legalisms. They wrote an artifical DOS compatibility test, encrypted it, and shielded it against debuggers. With the technology of the time, it took a hardware-based debugger to dig it out. (Today virtualization could have done the same thing.)
For a counter example, think back to the look'n'feel lawsuit of Apple vs Microsoft. That's a time when the/. crowd would have been on Microsoft's side, had it existed./.-precursors were on Microsoft's side. But of course in those days, Microsoft wasn't the Evil Empire, IBM was.
Sure, you can buy a dead or dying company and sue as if the damage had been done to you, but...
You just bought the company, and not only do you get the parts you wanted, you get the other parts, too. Just ask look into buying former/current property used for dry cleaning. Simpler yet, buy property with asbestos insulation.
Perhaps the new SCO bought the old SCOs damage, but they also bought the old SCOs actions wrt Open Source, including Christoph Hellwig's contributions, and all the implications thereof.
My other hope in this current chunk of mess, aside from SCO getting what it deserves, is that Hellwig doesn't suffer for any of this. Not to neglect the rest of the Open Source community, but it may well end up with Hellwig being at the eye of the storm.
I agree wholeheartedly with what you say about the days when Bell Labs and IBM Research did unfettered research.
Unfortunately, these companies were under "market attack" by companies like Dell, which does no technical research at all. They're also under "analyst attack" by Wall Street types measuring progress one quarter at a time, with little eye to the long term.
The logical consequence of this is that research gets "focused," unfortunately.
One *could* take messes like this to reinforce SCO's point - that the GPL is terrible, (terrorist?) and an excessive hindrance to business. Therefore the GPL should be found illegal and all of those copyrights reassigned as traditional copyrights to the correct entity appropriate to a derived work of Unix, namely SCO.
This is meant as tongue in cheek. Problem is, I suspect SCO could argue this in court with a straight face, and others like Microsoft would back them with an equally straight face.
Yeah, and unless you're at a University, DOD installation, or the like, you WON'T have access, because ISPs will only connect you to the Microsoft Internet.
Kiss your non-Windows connection goodbye, mine too.
Well, actually the Soyuz can't be *permanently* docked, or it wouldn't be very useful as an evacuation vehicle. How about 'long term'?
But there is another consideration. The Soyuz has a shelf-life, and they periodically have to change the thing out, anyway. Every so often a Soyuz mission will come up, and take the rescue capsule down, leaving their original transport as the new rescue capsule.
There's the old saying, "Power corrupts. Absolute power corrupts absolutely."
I once heard an interesting tweak on that, and perhaps more true than the original. "Power attracts the corruptible."
Perhaps Faramir really IS as pure as all that. Perhaps he never sought any greatness or position, only to do his best for his people. In that case, any station he has would be purely as a result of people under him pushing him up. Perhaps those of higher station yet were either born to it, or sought it, the latter implying that they are likely corruptible.
I've considered this idea more generally in the past... paying a 'disposal fee' up-front on new goods to pay for their end-of-life costs. There are two problems with this idea:
1: Technology changes, and those end-of-life costs are going to change, sometimes up, sometimes down. This in itself isn't a terrible problem, but it couples into problem 2.
2: Disposal escrow would wind up creating some huge lumps of money. IMHO, whenever there's a huge lump of money, there's also a class of people who will find a way to attach themselves to it and start sucking it dry. In other words, that lump will never survive to do what it was supposed to do - pay disposal costs. Relative to item 1, someone (from that class) will find a 'new technology' to handle disposal and use the fund to develop that new technolgy. Maybe it'll work, maybe not, but odds are that the point will have been to gain access to the money, not to develop technology. Let's presume that 50% of the time the technology falls through, and the money's gone. We're right back where we started, only with a broken promise and either an environmental mess or the need for another government bailout.
Offhand, I think it would be more fun to see William Hurt play Capt. James Tiberias Kirk in a new remake of Star Trek Classic. Can't think of anyone else more capable of Shatnering the role.
You're absolutely right, and for some reason I hadn't made that little connection in my mind between using GPG keys to sign packages or MD5s and the traditional web of trust.
But there's a problem here from the 'simple consumer' perspective. For the web of trust to really work well, you've got to join it and participate. I don't argue at all that that works well for the developers. But I can see a problem if 'simple consumers' join in.
Simple consumers won't participate well in a web of trust. Joe Sixpak will trust his friend Colin Compu-nerd, without checking on Colin's trust-path. Mike Modem is a friend of Joe and Colin, and trusts them. Before long you have a small pool of trust, completely disconnected from the real web of trust. One or more of those guys chooses to blindly trust some keys off of a website, and the others trust them, too.
To really work well, the web of trust needs members, not clients who feign membership to gain some capability or access. That's why I proposed some sort of key-publisher-with-votes, to allow non-participating clients. Forcing would-be clients to become feigning members weakens the web, too. Allowing them to remain clients allows the web to consist of true members, keeping it strong.
We can't all be Kevin Bacon.
The web of trust graphs are neat, but another neat thing would be an Oracle of Bacon, showing the trust hops that connect me to another person, or list of people.
Then the next point of failure becomes the keyservers. How do you know you imported a good key, and that the keyserver hadn't been compromised when you did it?
This probably would be no good as a way to sneak backdoors onto more than a few machines, since keys are usually stored once and used often. But it would be good to have some sort of key distribution and verification system. Imagine a key publisher having 7 peers, and where they carry same keys, requiring 5 to 7 matching signatures, and point a nasty finger at the odd one(s). More than two mismatching signatures and the system quits publishing keys.
Of course then the key publishers themselves then become a choke point for a DOS attack, of sorts. Make updates grind to a halt as a new exploit is emerging, widening the window to utilize it. But still, most keys are stored, and the voting fails only stop distribution and verification.
Thorny issues, part of why PKI is considered 'hard'. But at least my suggestion is reasonably decentralized (I didn't say how to get a new key into the system) and has publishers voting on the intersection of their published keys, not requiring every server to publish every key.
So don't cave into the temptation to buy the August release, unless you want both versions. Quoth Scottie, "Trick me once, shame on you. Trick me twice, shame on me." Anyone for three?
Slashdot Mirror
I guess I should have said, "The region of the Laffer Curve where increased econonomic activity will compensate overall tax revenue lost by reducing the tax rate." Reagan clearly wasn't on that part of the Laffer Curve, either. And perhaps that wasn't really the problem, it's that neither Reagan nor Bush II cut taxes while keeping spending under control.
Is it so they can drop support, or is it so they can drop the one reliable method of moving on-machine files between Windows and Linux?
FAT compatilibity is stunningly important to Linux. Assuming much Linux migration begins as dual-boot, FAT stitches the two sides of the box together. FAT is how we read/write those silly digital cameras across platforms. FAT is how we read/write those silly memory keys across platforms. FAT is how we read/write diminutive floppies across platforms.
Let's see DSC00072.jpg
That's 12345678.123
I was just futzing with this last night on my files that came from a Sony digital camera. If the patents truly are related to long filenames, looks like Sony might be covered, unless they support long directory names (Entering LFNs them using the UI of a digital camera would be a royal pain.) or unless the name gets one digit longer on my millionth picture, assuming the camera lasts that long and I insist on the ability to store every picture I take with it in one directory.
This is a spike. It'll last until the deficit catches up to us, presumably after November, 2004. I'm just trying to figure out when to make my stock moves. The economy isn't doing SO well that it'll put us onto the Laffer Curve.
That was the intent, and I suspect to some extent it still works that way. But I suspect that there's far more happening of a different sort with patents - and that's the creation of a club. Remember the phrase, "stand on the shoulders of giants?" Well, the shoulders are patented. If you want to stand on those shoulders and reach higher, you have to let the giant reach that high, too.
In essence, patents have created a club, and while you can still get a patent and make money, you probably can't disrupt an existing technology, because you need to license existing technology to make your patent work, and the most likely license term is to cross-license your technology back to the would-be disruptees. They can either take advantage of the technology, or you'll find that the license prevents you from disrupting their business - unless you're excessively lucky.
I recently heard about a guy with some sort of chemical/drug/food (forget which) patent that's running out. NONE of the industry has agreed to license it, they're just waiting for it to expire. In the meantime he's losing all of his development and attempted marketing money. Maybe he was asking absurd terms, maybe he deserved them, but the industry felt we could get along without the new product, the guy couldn't commercialize without more money than he had, so they could afford to wait.
Come to think of it, I've got a friend in the very same situation.
Which brings up a pet thought of mine...
What about a 'transfer station' in a highly elliptical orbit around both Earth and Moon? Agreed that a rendezvous with the transfer station would be every bit as energetic as getting to the Moon, but it could be done with something minimal like a Soyuz capsule, in terms of crew space. Once docked, the crew space of the transfer station could be much more spacious, maybe even with some equipment for science. Consumables would have been carried aboard the rendezvous capsule, with the crew. Upon return from the Moon, the lower mass capsule could again be used to get to more Earthly speeds.
Of course the same orbital mismatch would occur at the Moon, perhaps solved in a similar fashion. The transfer station would keep some sort of reusable LEM docked, except when being used on the Moon. That LEM would have the additional requirement of reaching lunar escape velocity in order to rendezvous with the transfer station.
Take the idea a little further - space stations in orbit around both Earth and Moon, and a transfer station orbiting both. Get from Earth to the Earth station, go from there to the transfer station, from transfer station to the Moon station, from there to the Moon. The essence is to get things down to minimal, efficient, single-purpose vehicles, and accelerate as little mass as possible at any given step.
Sure there have been security breaches, security breaches will always be with us. Anyone who can't accept that is probably a prime customer for Palladium, and deserves what they get.
The real issue here is that Debian and Gentoo were both forthcoming about the breaches. They both did the Right Thing. Not only that, but they've both collected forensics, and if not identifying culprits, are at least contributing to improving the security of the Linux community.
This is Real Security, as opposed to hiding the facts, and hoping nobody ever finds out.
Convenient leading choice of metaphor, thanks.
There are some who theorize (Sorry, no reference, I read this years (decades) ago in dead tree form.) that life began bootstrapped on clays.
Some clays at the microscopic scale carry electrostatic charges, and at that level can be somewhat self-assembling through sedimentation. So there is a micro-structure there. Now take the water that that microstructured clay is immersed in (We are talking about sedimentation, after all.) and put some organic impurities in it. Those impurities will tend to self-align on the clay face, drive by diffusion, micro-scale surface topology, and electrostatic charge. At some point you may have a sufficiently advanced scum on the clay surface that if a chance current breaks it off, it can 'do something' on it's on - a sort of proto-life.
This scenario is a case of a self-assembled unit that assembles something else.
On the serious side, before getting to the fantasy below...
There is a fuzzy dividing line between the macro and microscopic worlds. Various experiments have been able to move that line, AFAIK always upwards, but the line is still there. At some point, materials quit acting classically and start acting quantumly. Molecular assemblers are clearly below that line.
It's really hard to 'work your way down' to the molecular assembler level, because you're either classical or quantum. Even the experiments that elevate quantum effects into normally classical space are highly specific, and could not likely be extended to an assembler.
The chicken-and-egg problem mentioned elsewhere has a solution. Some 'proto-chicken' laid the first 'true-chicken' egg that hatched and became the first chicken. Rather obvious with a little thought.
The molecular assembler isn't so easy, and building the bridge is tough. I see two ways, either building really simple assemblers in bulk, in 'test tubes' somewhat as we build lipospheres, today, and using them to build more complex assemblers. Or building simple, gross molecular assemblers out of the smallest classical assemblers (STM on steriods?) we can make.
In either case, it's a bootstrap process, and by the time we've bootstrapped we may find that we're not exactly where we expected to be. Not necessarily better or worse, just different. Pre-bootstrap expectations are likely to be wrong.
A pet fantasy from my high school days (early 70's) was to build a smaller Waldo-style robot. (mimic my movements, at another place/scale) Use that Waldo to build a yet smaller one, and so on.
The final goal of this progression was to go walking down the grooves of my favorite vinyl LPs, carrying a micro-trowel and a micro-bucket of vinyl patch, and fix the really annoying clicks and pops by hand.
This was never a serious fantasy, but a fun thought. Not only did I not figure the needed size of the Waldo, beyond the 1-mil (25.4 uM) scale, I didn't figure the length of the virtual hike it would take for one LP.
Not only were vinyl LPs obsolete before nanotech (It didn't have that name, then.) truly approached, my old 'temporary, until I can afford better' AR-XA turntable outlasted the format.
Not to stretch a metaphor, but what does it mean that you can't stay in the eye of the storm by standing still or staying where you are? To stay in the nice weather you have to move with the storm - rather exactly.
Won't argue with you on this one. Assuming the damage is real, selling it gives Party B immediate revenue, allowing it to focus on its core business.
I guess there can be different ranges of 'ethical-ness' for Party C, and it might be questionable whether I'd invite ANY of them over for dinner, though I'd be more likely to invite your first group of Party C than your second group.
Of course developers should Listen to all feedback.
How much of that feedback they should Act on is an entirely different matter. It's probably more important to get some sort of demographic feel for their customer base, to know if it's some sort of lunatic fringe making the most noise, or if complaints are broad-based.
Perhaps the hardest challenge would be developing a game that can appeal both to a broad base, and to hard-core lunatic fans.
One can examine issues and take sides based on them, rather than on the players. Try a search on "aard" "microsoft" and "drdos" for more information. Years back, Dr. Dobbs ran a particularly condemning article on the AARD code. From what I could see, wasn't much 'alleged' about it, other than legalisms. They wrote an artifical DOS compatibility test, encrypted it, and shielded it against debuggers. With the technology of the time, it took a hardware-based debugger to dig it out. (Today virtualization could have done the same thing.)
/. crowd would have been on Microsoft's side, had it existed. /.-precursors were on Microsoft's side. But of course in those days, Microsoft wasn't the Evil Empire, IBM was.
For a counter example, think back to the look'n'feel lawsuit of Apple vs Microsoft. That's a time when the
How times change. How things stay the same.
Sure, you can buy a dead or dying company and sue as if the damage had been done to you, but...
You just bought the company, and not only do you get the parts you wanted, you get the other parts, too. Just ask look into buying former/current property used for dry cleaning. Simpler yet, buy property with asbestos insulation.
Perhaps the new SCO bought the old SCOs damage, but they also bought the old SCOs actions wrt Open Source, including Christoph Hellwig's contributions, and all the implications thereof.
My other hope in this current chunk of mess, aside from SCO getting what it deserves, is that Hellwig doesn't suffer for any of this. Not to neglect the rest of the Open Source community, but it may well end up with Hellwig being at the eye of the storm.
I agree wholeheartedly with what you say about the days when Bell Labs and IBM Research did unfettered research.
Unfortunately, these companies were under "market attack" by companies like Dell, which does no technical research at all. They're also under "analyst attack" by Wall Street types measuring progress one quarter at a time, with little eye to the long term.
The logical consequence of this is that research gets "focused," unfortunately.
One *could* take messes like this to reinforce SCO's point - that the GPL is terrible, (terrorist?) and an excessive hindrance to business. Therefore the GPL should be found illegal and all of those copyrights reassigned as traditional copyrights to the correct entity appropriate to a derived work of Unix, namely SCO.
This is meant as tongue in cheek. Problem is, I suspect SCO could argue this in court with a straight face, and others like Microsoft would back them with an equally straight face.
Yeah, and unless you're at a University, DOD installation, or the like, you WON'T have access, because ISPs will only connect you to the Microsoft Internet.
Kiss your non-Windows connection goodbye, mine too.
Well, actually the Soyuz can't be *permanently* docked, or it wouldn't be very useful as an evacuation vehicle. How about 'long term'?
But there is another consideration. The Soyuz has a shelf-life, and they periodically have to change the thing out, anyway. Every so often a Soyuz mission will come up, and take the rescue capsule down, leaving their original transport as the new rescue capsule.
Never read that one, so the phrase must be making the rounds.
There's the old saying, "Power corrupts. Absolute power corrupts absolutely."
I once heard an interesting tweak on that, and perhaps more true than the original. "Power attracts the corruptible."
Perhaps Faramir really IS as pure as all that. Perhaps he never sought any greatness or position, only to do his best for his people. In that case, any station he has would be purely as a result of people under him pushing him up. Perhaps those of higher station yet were either born to it, or sought it, the latter implying that they are likely corruptible.
I've considered this idea more generally in the past... paying a 'disposal fee' up-front on new goods to pay for their end-of-life costs. There are two problems with this idea:
1: Technology changes, and those end-of-life costs are going to change, sometimes up, sometimes down. This in itself isn't a terrible problem, but it couples into problem 2.
2: Disposal escrow would wind up creating some huge lumps of money. IMHO, whenever there's a huge lump of money, there's also a class of people who will find a way to attach themselves to it and start sucking it dry. In other words, that lump will never survive to do what it was supposed to do - pay disposal costs. Relative to item 1, someone (from that class) will find a 'new technology' to handle disposal and use the fund to develop that new technolgy. Maybe it'll work, maybe not, but odds are that the point will have been to gain access to the money, not to develop technology. Let's presume that 50% of the time the technology falls through, and the money's gone. We're right back where we started, only with a broken promise and either an environmental mess or the need for another government bailout.
Offhand, I think it would be more fun to see William Hurt play Capt. James Tiberias Kirk in a new remake of Star Trek Classic. Can't think of anyone else more capable of Shatnering the role.
You're absolutely right, and for some reason I hadn't made that little connection in my mind between using GPG keys to sign packages or MD5s and the traditional web of trust.
But there's a problem here from the 'simple consumer' perspective. For the web of trust to really work well, you've got to join it and participate. I don't argue at all that that works well for the developers. But I can see a problem if 'simple consumers' join in.
Simple consumers won't participate well in a web of trust. Joe Sixpak will trust his friend Colin Compu-nerd, without checking on Colin's trust-path. Mike Modem is a friend of Joe and Colin, and trusts them. Before long you have a small pool of trust, completely disconnected from the real web of trust. One or more of those guys chooses to blindly trust some keys off of a website, and the others trust them, too.
To really work well, the web of trust needs members, not clients who feign membership to gain some capability or access. That's why I proposed some sort of key-publisher-with-votes, to allow non-participating clients. Forcing would-be clients to become feigning members weakens the web, too. Allowing them to remain clients allows the web to consist of true members, keeping it strong.
We can't all be Kevin Bacon.
The web of trust graphs are neat, but another neat thing would be an Oracle of Bacon, showing the trust hops that connect me to another person, or list of people.
Then the next point of failure becomes the keyservers. How do you know you imported a good key, and that the keyserver hadn't been compromised when you did it?
This probably would be no good as a way to sneak backdoors onto more than a few machines, since keys are usually stored once and used often. But it would be good to have some sort of key distribution and verification system. Imagine a key publisher having 7 peers, and where they carry same keys, requiring 5 to 7 matching signatures, and point a nasty finger at the odd one(s). More than two mismatching signatures and the system quits publishing keys.
Of course then the key publishers themselves then become a choke point for a DOS attack, of sorts. Make updates grind to a halt as a new exploit is emerging, widening the window to utilize it. But still, most keys are stored, and the voting fails only stop distribution and verification.
Thorny issues, part of why PKI is considered 'hard'. But at least my suggestion is reasonably decentralized (I didn't say how to get a new key into the system) and has publishers voting on the intersection of their published keys, not requiring every server to publish every key.
So don't cave into the temptation to buy the August release, unless you want both versions. Quoth Scottie, "Trick me once, shame on you. Trick me twice, shame on me." Anyone for three?