Have you ever played with a 2-part pendulum? It's a chaotic system
Yes. My legs. And despite their chaotic oscillations they haven't
broken off just yet:)
This isn't a freshman physics pendulum, or a guitar
No, but I would submit that the guitar model is closer to an SE
than the suddenly stopped robot arm, and easier for most people to
relate to.
You have energy stored in waves in the ribbon...
Well, IMHO I had already answered why that wouldn't cause the
ribbon to break. Something I haven't seen you take into account
though is that the energy is being continuously reduced by friction.
The longer the cable, the more energy is required to overcome it and
reflect. Some of those chaotic oscillations just won't make it to the
other end at anywhere near their initial amplitudes. Wavefronts can't
coalesce if they can never coincide. Did I mention that this is a
very, very long cable?
it's not like there will be some easy "send the next one up
*now* to correct for the effects of the last one"
Well, you may be right that there may not be enough of an effect to
make it worth worrying about to that degree, but if you were
worrying about it, the fact is it would be easy to do that.
While you could attempt to model all of this
Well, yeah -- you could model the oscillation aspect pretty well by
building a Lunar SE first and inducing the oscillations yourself.
the equations are certainly not solvable
Which equations specifically? You sound like you're making it up.
Somehow people seem to think that if we have a material strong enough,
the rest will be easy
I agree with that. There's plenty of other things to worry about
that will make this project an engineering challenge -- and also why
it can be a superlative human achievement when it is completed. Don't
ask me to lose any sleep over the oscillation aspect though.
how do you deploy the thing - manufacture it all in geosynchronous
orbit
Well, now you're changing the subject, but I haven't heard anyone
claim that it would be manufactured anywhere in orbit.
The publishing industry has been around for a long long time, and
so they have built up a lead with:
a fairly functional set of content filters,
a mature
and tested set policies and procedures,
a large cache of resources
(e.g. image archives, reference materials),
connections with
experts, decision-makers, etc.
I'd say it's a bit premature to compare them just yet, because
Wikibooks is not even in its infancy, having not yet rediscovered and
incorporated the equivalent capabilities. That doesn't mean that
wikibooks as a community can't grow to produce quality. The
community itself hasn't fully taken shape yet, so to speak. A lot of
smart and genuinely helpful people still don't know about it or have a
good reason to participate in such a way that brings them back
constantly. You have to have a critical mass of quality to attract
quality. A culture of trolls and egos can only be defeated once
enough knowledgable, self-disciplined and materially uninvolved types
get wise to them -- and get them to help instead. It will take time
to defeat the forces of ignorance:)
In the meantime, there is nothing to stop one or more college
professors from writing a book and making it available there, and then
for another one to notice and announce, "I have found version 1.414 of
wikibook X to be sufficiently accurate that we can use it for our
class. A local copy is on server Y. I suggest you do not use any
later versions until I have looked at the diffs, but feel free to
correct any errors you find yourself." A student may add a graph of
lab data without offending any experts, I would think. Ideally, a graduate
class may also be able to keep up with the latest research if sometime
by the middle of the class a chapter is changed with the latest
results of a relevant experiment. No need to learn that fusion hasn't
broken even yet when new data reveals it already has.
Finally, let's not forget that the publishing industry hasn't been entirely unwilling to
produce mindless drivel from time to time. I know this isn't a
textbook example, but look at the mythic "immortality" ads in some
popular teen science magazines for instance. Sometimes, the articles
themselves are little more than an ad for some lame product. The
pseudoscience and crass commercialism that gets through the B&M
filters can also be a joke. But, it's a joke over which we have
virtually no influence.
I searched wiktionary for the etymology of the word planet. To make a long story short, it turns out that Latin "plane" means plain, clear or distinct, while Greek "planes" gives us "wanderer." OK, not much help there, the original choice of words wasn't itself all that well thought out, so we wouldn't be breaking some meaningful ancient tradition even if we called it a "rock" instead.
Why not settle the debate by saying that Pluto is a "Kuiper Belt Planet?" Kuiper Belt Planets can have their own definitions:) No need to burn all of last year's astronomy books.
...expecting to predict the exact wave form created by sending a
weight down a swinging pendulum, expecially one with a weight near the
middle in addition to at the end, is just impossible. It's only a
matter of time before you get it just slightly wrong and snap the
whip.
Expecting to predict the exact wave form created by plucking a
guitar string with your finger, especially one with another
finger near the middle in addition to at the end, is just
impossible. It's only a matter of time before you get it just slightly
wrong and snap the whip.:)
It sounds like you believe that resonance can become destructive at
any frequency, and that you have to worry about dampening every
frequency at every location at every microsecond. According to your
theory, all bridges would have fallen down by now due to the random
coalescence of wavefronts, and the ocean would have produced a wave
tall enough to reach the moon.
For the sake of argument, think of the SE as basically a huge guitar
string. It differs from your ukulele in that it has a very very long
period of vibration. That's the part that may be hard to intuitively
grasp. Think lunar tides, not middle C. Now, any guitar string has a
natural frequency that it likes to vibrate at, and you are not going
to build up enough amplitude/energy to break the string except at that
frequency.
Sending up a climber is like putting a finger somewhere on the
string; you will end up with a higher "pitch," but in doing so, you
have also cut down the maximum amplitude possible. The string just
won't play as loud with that finger on it as without it. With more
and more climbers, the ability to resonate destructively is not
increased as you seem to think (unless you deliberately synchronize
the peaks).
If you're still worried about oscillations, another experiment you
might want to try is to pluck a guitar string at 0.1% of its length to
see how much it oscillates. That's the part of the cable that's
subject to the weather, and represents an even tinier fraction of the
total mass.
The latency involved in communincating down the length of the cable
would make current proven realtime control technology useless.
Given that the period of vibration of the cable is measured in
hours, if it takes one whole second to send a signal to a
climber to tell it to slow down slightly, is that good enough to
appease the latency god, or are you wanting to dampen the vibrations
above 1 Hz as well?
The SE cable, in its current design, is kept stretched out, under a
[superlative] amount of tension. At eye level, the cable will feel as
hard as a road surface. Still, the engineers know that oscillation
will still occur, so they have a couple of good tactics they can use.
One is to keep track of where the peaks and valleys are and to time
the release of the next climber so that it produces oscillations
out of phase with the previous one, either cancelling them out
or at least not harmonizing into resonance. You can even adjust the
speed of each climber to keep them out of phase. Another tactic is
for the ground station to dampen the oscillations (think shock
absorber). Well, there is one other tactic and that's to wait for the
oscillations to die off as friction, but that's boring:)
Something that isn't in their current plans but which may be
incorporated at some point is a ground station with a device like a
tidal generator that can extract some energy from the longer period
oscillations. The energy can then be beamed back to the climbers.
... the problem with a space plane is [...that] it has to carry its own
fuel into orbit. That means you burn up most of your fuel boosting up
more fuel.
Unless it's one of those "launch from some other craft at high
altitude" ideas (but GP didn't allude to that).
GP seems to overlook that a spaceplane is also a "massive" point of
failure. You can have more than one spaceplane, but that doesn't
reduce the chances of one of them failing. Worse, if one of them does
fail, then the entire fleet is grounded anyway.
I agree with GP that the complete failure of the first stage
(single-cable) elevator would be more expensive than a single
spaceplane failure, depending on the cost overruns in either program.
But, neither the cable(s) nor the cargo of an SE have to travel x
times faster than the speed of sound, which more than evens the odds,
IHMO. The dangerously high speed is both a major source of risk for
the spaceplane, as well as its primary advantage for passenger travel
(to reduce the time of exposure to radiation).
The main advantage of a space elevator is that you dont
have to send up fuel; you only pay the energy cost necessary to boost
the payload.
Not only do you not have to send up fuel, you don't have to deposit
as much garbage in orbit that complicates future launches. It may
be possible to generate energy on the trip down, BTW, rather
than expending fuel to come back down.
I challenge you to come up with even one death certificate from anywhere that says "air pollution".
I don't think we can expect to get this sort of analysis from the coroner. A coroner will cite the last straw on the camel's back, not things that affect everyone continuously. We're all being bombarded by cosmic rays, for example, but the coroner reports the resulting cancer as the cause, not the cosmic rays.
We can expose lung cells to various contaminants in the laboratory and predict fairly reliably which elements will be harmful. There may not be a 1:1 correlation from lab to the real world, but that's not necessary to make the reasonable guess that someone is going to die from prolonged exposure or from indirect complications of respiratory illness.
I suppose we'll have to agree to disagree then (but at least thanks for listening and being polite about it).
Valid question, and we could also ask why the tobacco plant was
used instead of something edible, but the heat tolerance was just the
first phase. It was purely meant as a proof of concept that an
extremophilic gene of any kind could be introduced to a plant without
killing it due to any toxicity. They were successful. As a
double-bonus, you may be able to grow tobacco in Texas now, though you
may want to look up the property values in hell first.
That being said, and as I'd rather be insightful than funny, the
research could still be useful on the trip to Mars by allowing
more sunlight to hit the plants without killing them. This could reduce the shielding requirements, or possibly make the orbital farms maintenance-free and less likely to fail on the long trip. (Note that you could
send unmanned orbital farms out there ahead of time and watch the
plants grow with cameras. Then, if all goes well, send the crew on a
an express flight, and they won't have to load that down with as many
supplies.)
To grow stuff on Martian soil would seem to require some
precipitation as well, right? Conceivably, you could get algae to
grow directly in ice, but anything beyond that and you need to use
some sort of greenhouse system that melts the ice so that the plants
(and soil bacteria?) can grow.
pressurized hydrogen tanks hardly leak at all. but they aren't very energy dense
So it sounds like what you're saying is, that if you have enough space for the tanks and excess energy beyond the chemical battery capacity, you can store the energy as hydrogen, albeit with conversion losses.
The applications I'm thinking about are, for example,
an off-grid house which has excess solar energy because the inhabitants went on vacation or just had a long spell of sunny weather. The house can continue to store energy in underground H2/O2 tanks even after the batteries are topped off.
the hybrid car that is driving downhill and loses the energy that would otherwise charge the batteries because the batteries are already full.
As far as liquid hydrogen, I was thinking of maybe a tank of that, inside of a tank of ultra-compressed air to reduce the losses (and store additional energy as air pressure), but I don't know if it would be worth it for the house if there is plenty of space available anyway.
(Posting under the assumption that this topic is germane at all to this discussion...)
ID does not preclude the creation of life on earth by Alien life forms from another planet after 'Terra'forming the earth
Say that aliens landed and designed life on Earth. The value of a theory (to science) is in how accurately it makes a prediction. Now, whom does ID, if it is a theory, predict would have designed those Aliens?
The most important message at the 2005 WinHEC about Microsoft's
trusted computing effort, now known as Next Generation Secure
Computing Base (NGSCB), is that it is late and will not be included in
Windows Longhorn [...YET]
NGSCB is late? So what, Longshorn itself will be delivered late.
They have simply moved the dates, but that doesn't say that they've
changed their minds. You are trying to lull people into thinking that
"it's OK, this time it will be different" and they won't act
monopolistically, lock people into their business model, centralize
internet identities, break standards, etc. But the motive and the
opportunity is still there, as strong as ever, and you have given us
no reason to trust Megalosoft this time, especially with a scheme as
potential powerful as this one is.
In short, you are wrong that Vista in 2006 will use TPMs with
certs for DRM purposes [...YET]...Maybe never...
You really can't give a guarantee that it won't happen. That's
what's wrong with your argument. You're saying, it's OK for everyone
to go and wait inside of the simmering pot, because the heat won't be
turned up to boiling... yet.
Can energy be stored as hydrogen for a longer time period without losses than stored as battery charge?
(BTW, I don't know the answer -- batteries leak and so do hydrogen tanks -- I'm wondering if anyone's already come across this and knows of a big difference that would make hydrogen preferred for long term storage.)
So you are telling me over half the people that died in India last year died from air pollution?
Yes, if you can believe both of our sources at the same time.
My source is based on the results of the Indian government's study (did you look at the picture in the article I mentioned yet?) Now, if you say you believe some other figure rather than theirs, then you are free to do so, but you haven't convinced me of the reverse, especially since I don't know which source you used, the reliability of the source and the year of the census it is based on, etc.
You're not arguing that air pollution causes no deaths per year, are you?
you'll find that an
inactive version of it is already embedded in the Prescott line and
god knows how many other models.
Thanks for pointing that out. Keep in mind though, chipmakers do have
the ability to try out several different kinds of prototypes before
they decide which design they will want to build in large quantities.
So, while I will happily admit that it's possible that I
underestimated the most likely current level of integration, it
is with the caveat that that early article could have been based on a
prototype that was never built. The data in it was fairly reliable,
but still speculative. The most likely thing, IMO, is that TC has
pieces in both the CPU and the chipset. Come to think of it,
that NX bit feature that they were in such a hurry to come out with,
now seems better oriented at addressing the only possible hole left in
TC than to protect existing hardware.
My main point, which is still valid, was to warn people that it's
more tricky than opening up the case and hunting for a discrete TPM
chip with an obvious target painted over it. Now that the vendors
know that people are learning about what TC really is, I think they
may try to hide the fact on any labels, boxes, or webpages (so this is
a good time to make your own archive of some sites). They already
changed the name of the governing body from TCPA to TCG -- I wonder
what other shell games are afoot?
The link that I cited presents it as being some sort of new thing,
which I agree could be kind of misleading if at least part of the
capability is already in the CPU, but I think that the DRM aspect of
it (e.g. control of sound, video channels) necessarily has to have
some support in the mobo chipset that handle those functions.
In particular, what does "not a full-bore
implementation" mean???
That is a reference to a comment made by IBM itself some time
(months?) back in an apparent attempt to allay fears about the
strength of the feature (maybe someone else remembers it offhand?) I
searched high and low for the link but aren't able to find it anymore.
My running cynical theory is that the only thing they didn't implement
was the self-destruct-on-tampering feature, and then only because
their own testers kept ruining them:)
Ideally what you want is something that absorbs the rays not just scatters them
I wasn't being ideal about it, just pessimistically low-budget:)
If what you want is to try to absorb the rays, you need lots of
volume. Big bags of air. The whole assembly will look something like
a cluster of grapes.
So as not to annoy the bremstrahlung deity, the skin of the grapes
will have to be thin. Either the skin will be self-repairing, or, in
a large enough production, you will need staff constantly checking for
leaks. Or both. Even with no punctures, H2 will be leaking out of
the bags, and the vacuum of space will be doing its part to help. So,
you will likely bring other gases along as well... While we're at it,
maybe also some smaller, transparent water filled bags with algal
farms to process wastewater and generate oxygen. Those could be used
as a nice self-repairing backup system in case the solar panels or
other systems fail. That's what I would put on the outside of the
spacecraft.
On the inside, ideally speaking, we would want lots of space again.
So, the conclusion is that to do things ideally requires more bags and
more assembly time in LEO, and it's still not a perfect system.
If we had the technology to build something like the Monoliths, I
suspect we would have invented FTL travel...
Didn't you see the movie, or at least read the book (definitely
worth reading)? The monoliths weren't built by us, they were built by
the aliens. So all we have to do is find one, and (rather than
contacting the aliens with it, as we wouldn't trust them anyway...) we
could use it as shielding for our Mars trip:)
Just because you don't see a TPM on a motherboard pic doesn't mean
that the same functionality hasn't been integrated into the silicon of
another chip.
On the Intel 945G mobo, this is exactly what has happened.
There, the TPM functionality is inside the chipset that accompanies
the CPU. The chipset typically handles the interface to DRAM and
controls the flow of data to output devices, among other things. By
the time the MacTels roll out, the TPM will most likely not be a
separate chip anymore (to sibling: that's how the developer configuration and the
final configuration can be workalikes.)
Those of you who plan to be looking for a chip labelled "TPM" on
the board as a way of determining the truth of Apple's claims by that
time will be wasting time. The only way to know (for the moment) is
to look for a TCG-conformant chipset model instead. In the future, you will have no need to check because all Intel chipsets are to be TCG conformant. So, unless Apple is claiming they will use an older chipset, the most stringent DRM capability ever
released to the mass market under the bizarro term of "Trusted
Computing" will be in there.
Note that on the Cell processor, the TPM is already in the CPU
itself, with no external signals to tap into, though IBM claims it is
not a full-bore implementation. In the future, as they try to cram more
transistors into a smaller space, Intel may also integrate most of the
chipset (and the TPM along with it) into the CPU. AMD has already
integrated the memory interface into the CPU on some of its
processors, and has also jumped on the TCG bandwagon (either that or be run over by it), so it is
only a matter of time for them to add a TPM as well.
The only thing that the pictures can prove is which stage of TPM
integration is being used by some developers:) None of it should be
interpreted to mean that Apple will not have a TPM somewhere. Their claim is not credible, in my humble opinion.
Well you produce 3 million death certificates that say "air pollution" that are dated 2003 and I'll believe you.
OK, let me give it a try:
5 million premature deaths every year from air pollution in
India alone.
The same article claims 60-70% "of air pollution in India is due to motor vehicles"
Draw some reasonable assumptions about the likely cause-effect relationships based on looking at the carcinogenic content of the cloud of exhaust in the picture in the link.
Figure that 5 million *.60 = 3 million.
You didn't really want me to specifically mail you all of the death certificates in an attachment did you? You wouldn't read them anyway.
Doing the same number crunching for the rest of the world (Bangkok, Manila, Mexico City, etc.), I would say that there's enough of a margin for error in the total, that it shouldn't be necessary to generate controlled experiments to prove the case for each cause of death.
Not that I'm an especially big fan of nuclear power, I just think people would just be leaping out of the pebble-bed pan and into a coal fire.
(Language note: I hate to discourage your courageous use of semi-foreign terms, but you may be confusing the Latin de jure (typically used in a legal context to designate who holds the right to something) with the French term, correctly spelled du jour and meaning "of the day," what you probably intended.)
...equivalent to the kinetic energy of a baseball traveling at
approximately 100 mph... so even with the loss of energy in the
collision... there is a lot of energy left over to cause havoc in the... shielding.
Sounds disappointing for shielding, but on the propulsion end, that
sure makes a case for solar sails. Or, maybe we should call them
"solar sieves":)
Going back to the GP:
You often get bremsstrahlung ("braking radiation") - the single
particle is instead replaced with a shower of much more dangerous
particles.
For me, this conjured an image of the black monolith in
A.C. Clarke's "2001, A Space Odyssey" floating in space and then I
thought, what would happen if something really solid like that were
placed some distance away from the astronauts?
Now, I don't mean right next to them (this is not to give them an
urge to use hand tools on each other:) ) I mean something like a mile
away or so, creating the dangerous particles with a really solid wall
of, say, moonrock, but giving them a chance to disperse like shotgun
pellets over that distance. While some particles would still match
the incidence angle of our courageous space travelers, the distance
should cause most of the particles to spread out into trajectories
that miss. Even though GP says that magnetic fields won't work here,
I wonder if the particles could be strongly charged during the initial
collision and then passed through a contraption similar to a particle
accelerator but with the intent to detour or defocus the beam further
on its way out.
One catch to this is that keeping a large and heavy monolith
interposed while in orbit could be problematic, what with the sun appearing at many angles. Once en route to Mars
though, the biggest problem would be having to tow it along (assuming
there is no space elevator to just sling the whole assembly out there,
but IIRC they want a Mars mission before the SE is built).
Assuming that something catastrphic occurs, and you no longer have
any signed install disks, you can always disable TPM in the bios,
install the new system, then re-enable the bios.
If your BIOS allows the TPM to be disabled, how does an untrusted
installer program retrieve the trusted key that signs the OS that is
to be installed?
The TCG spec absolutely promises that the physical operator of a machine can reset the TPM, clear all keys...
And how does the user perform these things, with jumper pins? By tapping in morse code on the plastic sticker on the chip? (...Spitballs?)
The user will have to try to access this functionality through the BIOS or the OS. The TPM doesn't drive the video or keyboard. So, it's academic: if the BIOS writer decides not to pass the capability along to the user, then, what are the user's options? 1) Sit and obediently watch the ads 2) The landfill.
Now, of course, you're going to reply "but this wouldn't be compliant with the TCG spec! You're not claiming that they might *gasp* deviate from the spec, are you?" Two words: "Embrace and Extend." (And don't reply just to tell me that that's three words.)
When you read specs by self-interested industry groups, you can't forget that they enjoy defining words, such as "user" differently from how you would. They may conveniently assume that the "user" is the chip buyer, not the end user, as far as their own implementation is concerned. As ridiculous as it sounds, any vague terms in the spec are considered fair game, and that's the way they like it. That way, they can violate the spec and still insist that they are 100% compliant, whether it be Java, USB 2.0 or TCG.
But you have to have physical access to the box to insert a CD and reboot it. Some highly secure operating systems don't allow the CD tray to be opened on a running system by a non-privileged user, so they'd better have a lot of time if they don't want someone to notice....
Then, if the box needs the recovery CD because it crashed, your only choice may be to wip the entire disk and reinstall. Also, if someone's cheap DVD drive ate your only signed install CD, you may succeed in locking yourself out.:)
This is similar to the hard disk password that most people didn't set, because they were more afraid of themselves, that they would forget the password and lose their data than the off-chance that someone would try to steal their lame computer.
I know a lot more about the TPM than what is in any FAQ; I've read
every page of the spec and have written software.... okay?
If you are the expert you say you are, then it should be obvious to
you that the preceding attestation is insecure.:)
...there is no interest in general purpose PCs to stop them
booting other OS's.
Here, you are not supporting your argument on a technical basis
(e.g. Intel writes the BIOS, therefore Intel decides if Linux can
boot), but a social one (e.g. Intel likes Linux, really we do
*blink* *blink*).
Intel makes money off of Linux and they are never going to stop
making machines that can run it!
Is that a promise? Sign here. Well, that's not where I was going
with this, but I have no quarrel with that as a goal.
The problem with what you describe is you could patch the BIOS to
disable this check of the TPM
That depends on how picky the boot ROM wants to be about the BIOS
it checksums. If you are saying that in some current crop of systems it's not going to resort to being picky, that is something I
might be led to believe. But, that's something that may be changed in
the future, without a public announcement, if, say, MSFT requires it
from Dell as part of some pact. In fact, come to think of it, they
had announced a strategy to lock Linux out of PCs a while whileback.
And there are other restrictions that the BIOS may have to fulfill
before it can even start, but I won't get into those.
In short the TPM is not there to keep the machine from booting other
software....
I don't [have to] doubt your [presently] good intentions about that. MSFT,
who has bought Gator/Claria, and other spyware vendors will have their
own plans for it. They are also in the business to make money. I
won't describe the horrid things they are capable of doing with this
because I don't want to give them any ideas:)
...That does have DRM implications, but the details are totally
different from the misleading picture that you described.
You are once again focusing on what the most immediate systems do,
while ignoring the capabilities other people will use it for,
especially a couple of years from now. I'm more interested in what
the full capabilities are, and where this is going. Are we
unwittingly enabling a 1984 type of future for ourselves? If
so, maybe the name of Sapiens Sapiens should be reconsidered.
Well, aren't we all today at the mercy of the vendor
To a certain extent, but it was never as restrictive and
potentially annoying as what they are planning. Soon, I predict it
will be illegal to sell computers without a TPM (and it will be
on-chip, invisible and unremovable, like the Cell). When someone
richer than God in Redmound proclaims "Consumers have too many
choices," you see this facile conclusion and you know that it
basically justifies some upcoming scheme to reduce your choices in a
self-serving way.
once the file is on your disk how can it force the OS to prevent
access from unauthorized programs
It doesn't have to. A DRM file would be stored encrypted. The
only way to decrypt it is to use the trusted application that it is
mated to. Such an application typically won't offer the option to
decrypt a DRM file. Therefore, if it wants to quit playing a song
after 5 listens, they can enforce that. Without the key from the
vendor to unlock it, a DRM file is just a pile of gibberish. I'm not
saying the OS will stop you from listening to an unencrypted file you
already have. That decision is up to the vendor. There are a lot of
decisions that can be made by the vendor (such as encrypting the
filesystem) that can dramatically impact the freedoms and flexibility
you have come to expect.
In your case, you are not running the full monty yet (a
TCPA-compliant Longshorn), which is why it seems so harmless. I'm not
as optimistic as you are about what's coming down the pike. To me, Trusted Computing is like having an M-1 tank on your doorstep. Sure, it's going to be fairly harmless if there are no keys to open it, but the keys will come someday, and you won't be allowed to hold them.
You claim:
It can hold keys, and sign and encrypt data with them. It's
completely passive. It never takes control of your system...
Sorry, there's a little bit more to it, unfortunately. From the
TCG's own FAQ,
... security processes... are protected through the secure TCG
subsystem.
Access to data and secrets in a platform could be denied if
the boot sequence is not as expected...
Features include... attestation of machine configuration when booted...
It sounds simple enough, but there is a whole realm of implications
that will someday come home to roost.
(Beware when reading the TCG's own FAQ, by the way, as they adopt a deceiving "don't blame us, we're not
the ones pulling the trigger" position. So, they gloss over some of
the juicy possibilities a BIOS writer or an application writer will
likely exploit from the technical specs.)
To begin with, the first application that boots up, typically the
BIOS (probably UEFI but any other choice really), if written to do
so can refuse to allow any application to start which isn't signed
by one of the keys securely stored in the TPM. The BIOS will check
the TPM for a matching key for the OS, and if it matches, will allow
it to start. Conversely, if the key doesn't match (for example, a
bootleg OS), the BIOS can just stop right there. Keep in mind, this
is the BIOS handling this, not the TPM, but, unlike even the M-1 tank,
there is no way to tamper with the TPM to change the keys.
Now, once a trusted OS is able to start, it can decide pretty much
autocratically what other applications can start, once again using the
keys locked down by the TPM to check if they are legit or not. So,
programmatically, the TPM doesn't make the decision to lock you out of
using non-vendor applications, but it's just as well as if it
did, because the OS writer can easily use the TPM's secure,
untamperable storage to enforce it. (Note that the motherboard supplier can cooperate with
the OS writer to initialize the TPM with the appropriate keys right
out of the factory (if they wanted to). It's irrelevant if there are
no keys in there right now. The tank is still there, pointing at your
door, waiting for its keys to arrive.)
Other applications, if they are also signed by the TPM, may be
granted the privilege (by the OS) to start and, specifically, to lock
down data, such as video, in order to provide DRM functionality. If
that decision is made, there is no way you will see that video through
any other application unless the application governing the data allows
otherwise. That data can basically be owned entirely by the
application vendor, not you (as different from what the TCG claims,
because no one's going to enjoy watching encrypted video gibberish.
You can technically "own" the gibberish, but you still can't watch the
video...). You may have a choice to delete a video, for example, but
not to view it unless that vendor allows it. It is a backdoor way of
implementing the media (DVD, CD, etc.) equivalent of the broadcast
flag, if the app writer and OS vendor cooperate to that
effect.
Unsigned applications may be allowed to start too, and the TCG
spec says that this is in the "user's" control, but let's face it,
it's really in the OS vendor's control because they control the
machine all the way from bootup. There isn't a little switch on the
TPM chip to allow you to override your OS' choice in the matter.
Still, it's possible that
Extending it even further, is there really any area left that could not be interpreted as competition, when MSFT's patent portfolio is so large and vague?
They "make" software, documents, hardware and services, and they also "produce" intellectual property (whether real or imagined). If the next company also produces any of these things, they may be a target. That doesn't leave much, especially if the judge lets their lawyers play fast and loose with the wording. And, if it's ruled not to be competition, MSFT may decide to enter that market and then sue, because there is no restriction on them competing with their former employee (not very fair, is it?)
If non-competes are ever justified, they should be even less justified for convicted monopolists.
Slashdot Mirror
Have you ever played with a 2-part pendulum? It's a chaotic system
Yes. My legs. And despite their chaotic oscillations they haven't broken off just yet :)
This isn't a freshman physics pendulum, or a guitar
No, but I would submit that the guitar model is closer to an SE than the suddenly stopped robot arm, and easier for most people to relate to.
You have energy stored in waves in the ribbon...
Well, IMHO I had already answered why that wouldn't cause the ribbon to break. Something I haven't seen you take into account though is that the energy is being continuously reduced by friction. The longer the cable, the more energy is required to overcome it and reflect. Some of those chaotic oscillations just won't make it to the other end at anywhere near their initial amplitudes. Wavefronts can't coalesce if they can never coincide. Did I mention that this is a very, very long cable?
it's not like there will be some easy "send the next one up *now* to correct for the effects of the last one"
Well, you may be right that there may not be enough of an effect to make it worth worrying about to that degree, but if you were worrying about it, the fact is it would be easy to do that.
While you could attempt to model all of this
Well, yeah -- you could model the oscillation aspect pretty well by building a Lunar SE first and inducing the oscillations yourself.
the equations are certainly not solvable
Which equations specifically? You sound like you're making it up.
Somehow people seem to think that if we have a material strong enough, the rest will be easy
I agree with that. There's plenty of other things to worry about that will make this project an engineering challenge -- and also why it can be a superlative human achievement when it is completed. Don't ask me to lose any sleep over the oscillation aspect though.
how do you deploy the thing - manufacture it all in geosynchronous orbit
Well, now you're changing the subject, but I haven't heard anyone claim that it would be manufactured anywhere in orbit.
The publishing industry has been around for a long long time, and so they have built up a lead with:
I'd say it's a bit premature to compare them just yet, because Wikibooks is not even in its infancy, having not yet rediscovered and incorporated the equivalent capabilities. That doesn't mean that wikibooks as a community can't grow to produce quality. The community itself hasn't fully taken shape yet, so to speak. A lot of smart and genuinely helpful people still don't know about it or have a good reason to participate in such a way that brings them back constantly. You have to have a critical mass of quality to attract quality. A culture of trolls and egos can only be defeated once enough knowledgable, self-disciplined and materially uninvolved types get wise to them -- and get them to help instead. It will take time to defeat the forces of ignorance :)
In the meantime, there is nothing to stop one or more college professors from writing a book and making it available there, and then for another one to notice and announce, "I have found version 1.414 of wikibook X to be sufficiently accurate that we can use it for our class. A local copy is on server Y. I suggest you do not use any later versions until I have looked at the diffs, but feel free to correct any errors you find yourself." A student may add a graph of lab data without offending any experts, I would think. Ideally, a graduate class may also be able to keep up with the latest research if sometime by the middle of the class a chapter is changed with the latest results of a relevant experiment. No need to learn that fusion hasn't broken even yet when new data reveals it already has.
Finally, let's not forget that the publishing industry hasn't been entirely unwilling to produce mindless drivel from time to time. I know this isn't a textbook example, but look at the mythic "immortality" ads in some popular teen science magazines for instance. Sometimes, the articles themselves are little more than an ad for some lame product. The pseudoscience and crass commercialism that gets through the B&M filters can also be a joke. But, it's a joke over which we have virtually no influence.
I searched wiktionary for the etymology of the word planet. To make a long story short, it turns out that Latin "plane" means plain, clear or distinct, while Greek "planes" gives us "wanderer." OK, not much help there, the original choice of words wasn't itself all that well thought out, so we wouldn't be breaking some meaningful ancient tradition even if we called it a "rock" instead.
Why not settle the debate by saying that Pluto is a "Kuiper Belt Planet?" Kuiper Belt Planets can have their own definitions :) No need to burn all of last year's astronomy books.
So, we could have:
Expecting to predict the exact wave form created by plucking a guitar string with your finger, especially one with another finger near the middle in addition to at the end, is just impossible. It's only a matter of time before you get it just slightly wrong and snap the whip. :)
It sounds like you believe that resonance can become destructive at any frequency, and that you have to worry about dampening every frequency at every location at every microsecond. According to your theory, all bridges would have fallen down by now due to the random coalescence of wavefronts, and the ocean would have produced a wave tall enough to reach the moon.
For the sake of argument, think of the SE as basically a huge guitar string. It differs from your ukulele in that it has a very very long period of vibration. That's the part that may be hard to intuitively grasp. Think lunar tides, not middle C. Now, any guitar string has a natural frequency that it likes to vibrate at, and you are not going to build up enough amplitude/energy to break the string except at that frequency.
Sending up a climber is like putting a finger somewhere on the string; you will end up with a higher "pitch," but in doing so, you have also cut down the maximum amplitude possible. The string just won't play as loud with that finger on it as without it. With more and more climbers, the ability to resonate destructively is not increased as you seem to think (unless you deliberately synchronize the peaks).
If you're still worried about oscillations, another experiment you might want to try is to pluck a guitar string at 0.1% of its length to see how much it oscillates. That's the part of the cable that's subject to the weather, and represents an even tinier fraction of the total mass.
Given that the period of vibration of the cable is measured in hours, if it takes one whole second to send a signal to a climber to tell it to slow down slightly, is that good enough to appease the latency god, or are you wanting to dampen the vibrations above 1 Hz as well?
The SE cable, in its current design, is kept stretched out, under a [superlative] amount of tension. At eye level, the cable will feel as hard as a road surface. Still, the engineers know that oscillation will still occur, so they have a couple of good tactics they can use. One is to keep track of where the peaks and valleys are and to time the release of the next climber so that it produces oscillations out of phase with the previous one, either cancelling them out or at least not harmonizing into resonance. You can even adjust the speed of each climber to keep them out of phase. Another tactic is for the ground station to dampen the oscillations (think shock absorber). Well, there is one other tactic and that's to wait for the oscillations to die off as friction, but that's boring :)
Something that isn't in their current plans but which may be incorporated at some point is a ground station with a device like a tidal generator that can extract some energy from the longer period oscillations. The energy can then be beamed back to the climbers.
Unless it's one of those "launch from some other craft at high altitude" ideas (but GP didn't allude to that).
GP seems to overlook that a spaceplane is also a "massive" point of failure. You can have more than one spaceplane, but that doesn't reduce the chances of one of them failing. Worse, if one of them does fail, then the entire fleet is grounded anyway.
I agree with GP that the complete failure of the first stage (single-cable) elevator would be more expensive than a single spaceplane failure, depending on the cost overruns in either program. But, neither the cable(s) nor the cargo of an SE have to travel x times faster than the speed of sound, which more than evens the odds, IHMO. The dangerously high speed is both a major source of risk for the spaceplane, as well as its primary advantage for passenger travel (to reduce the time of exposure to radiation).
Not only do you not have to send up fuel, you don't have to deposit as much garbage in orbit that complicates future launches. It may be possible to generate energy on the trip down, BTW, rather than expending fuel to come back down.
I don't think we can expect to get this sort of analysis from the coroner. A coroner will cite the last straw on the camel's back, not things that affect everyone continuously. We're all being bombarded by cosmic rays, for example, but the coroner reports the resulting cancer as the cause, not the cosmic rays.
We can expose lung cells to various contaminants in the laboratory and predict fairly reliably which elements will be harmful. There may not be a 1:1 correlation from lab to the real world, but that's not necessary to make the reasonable guess that someone is going to die from prolonged exposure or from indirect complications of respiratory illness.
I suppose we'll have to agree to disagree then (but at least thanks for listening and being polite about it).
Valid question, and we could also ask why the tobacco plant was used instead of something edible, but the heat tolerance was just the first phase. It was purely meant as a proof of concept that an extremophilic gene of any kind could be introduced to a plant without killing it due to any toxicity. They were successful. As a double-bonus, you may be able to grow tobacco in Texas now, though you may want to look up the property values in hell first.
That being said, and as I'd rather be insightful than funny, the research could still be useful on the trip to Mars by allowing more sunlight to hit the plants without killing them. This could reduce the shielding requirements, or possibly make the orbital farms maintenance-free and less likely to fail on the long trip. (Note that you could send unmanned orbital farms out there ahead of time and watch the plants grow with cameras. Then, if all goes well, send the crew on a an express flight, and they won't have to load that down with as many supplies.)
To grow stuff on Martian soil would seem to require some precipitation as well, right? Conceivably, you could get algae to grow directly in ice, but anything beyond that and you need to use some sort of greenhouse system that melts the ice so that the plants (and soil bacteria?) can grow.
So it sounds like what you're saying is, that if you have enough space for the tanks and excess energy beyond the chemical battery capacity, you can store the energy as hydrogen, albeit with conversion losses.
The applications I'm thinking about are, for example,
As far as liquid hydrogen, I was thinking of maybe a tank of that, inside of a tank of ultra-compressed air to reduce the losses (and store additional energy as air pressure), but I don't know if it would be worth it for the house if there is plenty of space available anyway.
(Posting under the assumption that this topic is germane at all to this discussion...)
ID does not preclude the creation of life on earth by Alien life forms from another planet after 'Terra'forming the earth
Say that aliens landed and designed life on Earth. The value of a theory (to science) is in how accurately it makes a prediction. Now, whom does ID, if it is a theory, predict would have designed those Aliens?
The most important message at the 2005 WinHEC about Microsoft's trusted computing effort, now known as Next Generation Secure Computing Base (NGSCB), is that it is late and will not be included in Windows Longhorn [...YET]
NGSCB is late? So what, Longshorn itself will be delivered late. They have simply moved the dates, but that doesn't say that they've changed their minds. You are trying to lull people into thinking that "it's OK, this time it will be different" and they won't act monopolistically, lock people into their business model, centralize internet identities, break standards, etc. But the motive and the opportunity is still there, as strong as ever, and you have given us no reason to trust Megalosoft this time, especially with a scheme as potential powerful as this one is.
In short, you are wrong that Vista in 2006 will use TPMs with certs for DRM purposes [...YET] ...Maybe never...
You really can't give a guarantee that it won't happen. That's what's wrong with your argument. You're saying, it's OK for everyone to go and wait inside of the simmering pot, because the heat won't be turned up to boiling... yet.
Science question:
Can energy be stored as hydrogen for a longer time period without losses than stored as battery charge?
(BTW, I don't know the answer -- batteries leak and so do hydrogen tanks -- I'm wondering if anyone's already come across this and knows of a big difference that would make hydrogen preferred for long term storage.)
So you are telling me over half the people that died in India last year died from air pollution?
Yes, if you can believe both of our sources at the same time.
My source is based on the results of the Indian government's study (did you look at the picture in the article I mentioned yet?) Now, if you say you believe some other figure rather than theirs, then you are free to do so, but you haven't convinced me of the reverse, especially since I don't know which source you used, the reliability of the source and the year of the census it is based on, etc.
You're not arguing that air pollution causes no deaths per year, are you?
Thanks for pointing that out. Keep in mind though, chipmakers do have the ability to try out several different kinds of prototypes before they decide which design they will want to build in large quantities. So, while I will happily admit that it's possible that I underestimated the most likely current level of integration, it is with the caveat that that early article could have been based on a prototype that was never built. The data in it was fairly reliable, but still speculative. The most likely thing, IMO, is that TC has pieces in both the CPU and the chipset. Come to think of it, that NX bit feature that they were in such a hurry to come out with, now seems better oriented at addressing the only possible hole left in TC than to protect existing hardware.
My main point, which is still valid, was to warn people that it's more tricky than opening up the case and hunting for a discrete TPM chip with an obvious target painted over it. Now that the vendors know that people are learning about what TC really is, I think they may try to hide the fact on any labels, boxes, or webpages (so this is a good time to make your own archive of some sites). They already changed the name of the governing body from TCPA to TCG -- I wonder what other shell games are afoot?
The link that I cited presents it as being some sort of new thing, which I agree could be kind of misleading if at least part of the capability is already in the CPU, but I think that the DRM aspect of it (e.g. control of sound, video channels) necessarily has to have some support in the mobo chipset that handle those functions.
In particular, what does "not a full-bore implementation" mean???
That is a reference to a comment made by IBM itself some time (months?) back in an apparent attempt to allay fears about the strength of the feature (maybe someone else remembers it offhand?) I searched high and low for the link but aren't able to find it anymore. My running cynical theory is that the only thing they didn't implement was the self-destruct-on-tampering feature, and then only because their own testers kept ruining them :)
I wasn't being ideal about it, just pessimistically low-budget :)
If what you want is to try to absorb the rays, you need lots of volume. Big bags of air. The whole assembly will look something like a cluster of grapes.
So as not to annoy the bremstrahlung deity, the skin of the grapes will have to be thin. Either the skin will be self-repairing, or, in a large enough production, you will need staff constantly checking for leaks. Or both. Even with no punctures, H2 will be leaking out of the bags, and the vacuum of space will be doing its part to help. So, you will likely bring other gases along as well... While we're at it, maybe also some smaller, transparent water filled bags with algal farms to process wastewater and generate oxygen. Those could be used as a nice self-repairing backup system in case the solar panels or other systems fail. That's what I would put on the outside of the spacecraft.
On the inside, ideally speaking, we would want lots of space again. So, the conclusion is that to do things ideally requires more bags and more assembly time in LEO, and it's still not a perfect system.
If we had the technology to build something like the Monoliths, I suspect we would have invented FTL travel...
Didn't you see the movie, or at least read the book (definitely worth reading)? The monoliths weren't built by us, they were built by the aliens. So all we have to do is find one, and (rather than contacting the aliens with it, as we wouldn't trust them anyway...) we could use it as shielding for our Mars trip :)
Just because you don't see a TPM on a motherboard pic doesn't mean that the same functionality hasn't been integrated into the silicon of another chip.
On the Intel 945G mobo, this is exactly what has happened.
There, the TPM functionality is inside the chipset that accompanies the CPU. The chipset typically handles the interface to DRAM and controls the flow of data to output devices, among other things. By the time the MacTels roll out, the TPM will most likely not be a separate chip anymore (to sibling: that's how the developer configuration and the final configuration can be workalikes.)
Those of you who plan to be looking for a chip labelled "TPM" on the board as a way of determining the truth of Apple's claims by that time will be wasting time. The only way to know (for the moment) is to look for a TCG-conformant chipset model instead. In the future, you will have no need to check because all Intel chipsets are to be TCG conformant. So, unless Apple is claiming they will use an older chipset, the most stringent DRM capability ever released to the mass market under the bizarro term of "Trusted Computing" will be in there.
Note that on the Cell processor, the TPM is already in the CPU itself, with no external signals to tap into, though IBM claims it is not a full-bore implementation. In the future, as they try to cram more transistors into a smaller space, Intel may also integrate most of the chipset (and the TPM along with it) into the CPU. AMD has already integrated the memory interface into the CPU on some of its processors, and has also jumped on the TCG bandwagon (either that or be run over by it), so it is only a matter of time for them to add a TPM as well.
The only thing that the pictures can prove is which stage of TPM integration is being used by some developers :) None of it should be
interpreted to mean that Apple will not have a TPM somewhere. Their claim is not credible, in my humble opinion.
OK, let me give it a try:
You didn't really want me to specifically mail you all of the death certificates in an attachment did you? You wouldn't read them anyway.
Doing the same number crunching for the rest of the world (Bangkok, Manila, Mexico City, etc.), I would say that there's enough of a margin for error in the total, that it shouldn't be necessary to generate controlled experiments to prove the case for each cause of death.
Not that I'm an especially big fan of nuclear power, I just think people would just be leaping out of the pebble-bed pan and into a coal fire.
(Language note: I hate to discourage your courageous use of semi-foreign terms, but you may be confusing the Latin de jure (typically used in a legal context to designate who holds the right to something) with the French term, correctly spelled du jour and meaning "of the day," what you probably intended.)
Sounds disappointing for shielding, but on the propulsion end, that sure makes a case for solar sails. Or, maybe we should call them "solar sieves" :)
Going back to the GP:
For me, this conjured an image of the black monolith in A.C. Clarke's "2001, A Space Odyssey" floating in space and then I thought, what would happen if something really solid like that were placed some distance away from the astronauts?
Now, I don't mean right next to them (this is not to give them an urge to use hand tools on each other :) ) I mean something like a mile
away or so, creating the dangerous particles with a really solid wall
of, say, moonrock, but giving them a chance to disperse like shotgun
pellets over that distance. While some particles would still match
the incidence angle of our courageous space travelers, the distance
should cause most of the particles to spread out into trajectories
that miss. Even though GP says that magnetic fields won't work here,
I wonder if the particles could be strongly charged during the initial
collision and then passed through a contraption similar to a particle
accelerator but with the intent to detour or defocus the beam further
on its way out.
One catch to this is that keeping a large and heavy monolith interposed while in orbit could be problematic, what with the sun appearing at many angles. Once en route to Mars though, the biggest problem would be having to tow it along (assuming there is no space elevator to just sling the whole assembly out there, but IIRC they want a Mars mission before the SE is built).
Assuming that something catastrphic occurs, and you no longer have any signed install disks, you can always disable TPM in the bios, install the new system, then re-enable the bios.
If your BIOS allows the TPM to be disabled, how does an untrusted installer program retrieve the trusted key that signs the OS that is to be installed?
The TCG spec absolutely promises that the physical operator of a machine can reset the TPM, clear all keys...
And how does the user perform these things, with jumper pins? By tapping in morse code on the plastic sticker on the chip? (...Spitballs?)
The user will have to try to access this functionality through the BIOS or the OS. The TPM doesn't drive the video or keyboard. So, it's academic: if the BIOS writer decides not to pass the capability along to the user, then, what are the user's options? 1) Sit and obediently watch the ads 2) The landfill.
Now, of course, you're going to reply "but this wouldn't be compliant with the TCG spec! You're not claiming that they might *gasp* deviate from the spec, are you?" Two words: "Embrace and Extend." (And don't reply just to tell me that that's three words.)
When you read specs by self-interested industry groups, you can't forget that they enjoy defining words, such as "user" differently from how you would. They may conveniently assume that the "user" is the chip buyer, not the end user, as far as their own implementation is concerned. As ridiculous as it sounds, any vague terms in the spec are considered fair game, and that's the way they like it. That way, they can violate the spec and still insist that they are 100% compliant, whether it be Java, USB 2.0 or TCG.
it can help deflect hack attempts as well
But you have to have physical access to the box to insert a CD and reboot it. Some highly secure operating systems don't allow the CD tray to be opened on a running system by a non-privileged user, so they'd better have a lot of time if they don't want someone to notice....
Then, if the box needs the recovery CD because it crashed, your only choice may be to wip the entire disk and reinstall. Also, if someone's cheap DVD drive ate your only signed install CD, you may succeed in locking yourself out. :)
This is similar to the hard disk password that most people didn't set, because they were more afraid of themselves, that they would forget the password and lose their data than the off-chance that someone would try to steal their lame computer.
If you are the expert you say you are, then it should be obvious to you that the preceding attestation is insecure. :)
Here, you are not supporting your argument on a technical basis (e.g. Intel writes the BIOS, therefore Intel decides if Linux can boot), but a social one (e.g. Intel likes Linux, really we do *blink* *blink*).
Is that a promise? Sign here. Well, that's not where I was going with this, but I have no quarrel with that as a goal.
That depends on how picky the boot ROM wants to be about the BIOS it checksums. If you are saying that in some current crop of systems it's not going to resort to being picky, that is something I might be led to believe. But, that's something that may be changed in the future, without a public announcement, if, say, MSFT requires it from Dell as part of some pact. In fact, come to think of it, they had announced a strategy to lock Linux out of PCs a while while back.
And there are other restrictions that the BIOS may have to fulfill before it can even start, but I won't get into those.
I don't [have to] doubt your [presently] good intentions about that. MSFT, who has bought Gator/Claria, and other spyware vendors will have their own plans for it. They are also in the business to make money. I won't describe the horrid things they are capable of doing with this because I don't want to give them any ideas :)
You are once again focusing on what the most immediate systems do, while ignoring the capabilities other people will use it for, especially a couple of years from now. I'm more interested in what the full capabilities are, and where this is going. Are we unwittingly enabling a 1984 type of future for ourselves? If so, maybe the name of Sapiens Sapiens should be reconsidered.
To a certain extent, but it was never as restrictive and potentially annoying as what they are planning. Soon, I predict it will be illegal to sell computers without a TPM (and it will be on-chip, invisible and unremovable, like the Cell). When someone richer than God in Redmound proclaims "Consumers have too many choices," you see this facile conclusion and you know that it basically justifies some upcoming scheme to reduce your choices in a self-serving way.
once the file is on your disk how can it force the OS to prevent access from unauthorized programs
It doesn't have to. A DRM file would be stored encrypted. The only way to decrypt it is to use the trusted application that it is mated to. Such an application typically won't offer the option to decrypt a DRM file. Therefore, if it wants to quit playing a song after 5 listens, they can enforce that. Without the key from the vendor to unlock it, a DRM file is just a pile of gibberish. I'm not saying the OS will stop you from listening to an unencrypted file you already have. That decision is up to the vendor. There are a lot of decisions that can be made by the vendor (such as encrypting the filesystem) that can dramatically impact the freedoms and flexibility you have come to expect.
In your case, you are not running the full monty yet (a TCPA-compliant Longshorn), which is why it seems so harmless. I'm not as optimistic as you are about what's coming down the pike. To me, Trusted Computing is like having an M-1 tank on your doorstep. Sure, it's going to be fairly harmless if there are no keys to open it, but the keys will come someday, and you won't be allowed to hold them.
You claim:
Sorry, there's a little bit more to it, unfortunately. From the TCG's own FAQ,
It sounds simple enough, but there is a whole realm of implications that will someday come home to roost.
(Beware when reading the TCG's own FAQ, by the way, as they adopt a deceiving "don't blame us, we're not the ones pulling the trigger" position. So, they gloss over some of the juicy possibilities a BIOS writer or an application writer will likely exploit from the technical specs.)
To begin with, the first application that boots up, typically the BIOS (probably UEFI but any other choice really), if written to do so can refuse to allow any application to start which isn't signed by one of the keys securely stored in the TPM. The BIOS will check the TPM for a matching key for the OS, and if it matches, will allow it to start. Conversely, if the key doesn't match (for example, a bootleg OS), the BIOS can just stop right there. Keep in mind, this is the BIOS handling this, not the TPM, but, unlike even the M-1 tank, there is no way to tamper with the TPM to change the keys.
Now, once a trusted OS is able to start, it can decide pretty much autocratically what other applications can start, once again using the keys locked down by the TPM to check if they are legit or not. So, programmatically, the TPM doesn't make the decision to lock you out of using non-vendor applications, but it's just as well as if it did, because the OS writer can easily use the TPM's secure, untamperable storage to enforce it. (Note that the motherboard supplier can cooperate with the OS writer to initialize the TPM with the appropriate keys right out of the factory (if they wanted to). It's irrelevant if there are no keys in there right now. The tank is still there, pointing at your door, waiting for its keys to arrive.)
Other applications, if they are also signed by the TPM, may be granted the privilege (by the OS) to start and, specifically, to lock down data, such as video, in order to provide DRM functionality. If that decision is made, there is no way you will see that video through any other application unless the application governing the data allows otherwise. That data can basically be owned entirely by the application vendor, not you (as different from what the TCG claims, because no one's going to enjoy watching encrypted video gibberish. You can technically "own" the gibberish, but you still can't watch the video...). You may have a choice to delete a video, for example, but not to view it unless that vendor allows it. It is a backdoor way of implementing the media (DVD, CD, etc.) equivalent of the broadcast flag, if the app writer and OS vendor cooperate to that effect.
Unsigned applications may be allowed to start too, and the TCG spec says that this is in the "user's" control, but let's face it, it's really in the OS vendor's control because they control the machine all the way from bootup. There isn't a little switch on the TPM chip to allow you to override your OS' choice in the matter. Still, it's possible that
Extending it even further, is there really any area left that could not be interpreted as competition, when MSFT's patent portfolio is so large and vague?
They "make" software, documents, hardware and services, and they also "produce" intellectual property (whether real or imagined). If the next company also produces any of these things, they may be a target. That doesn't leave much, especially if the judge lets their lawyers play fast and loose with the wording. And, if it's ruled not to be competition, MSFT may decide to enter that market and then sue, because there is no restriction on them competing with their former employee (not very fair, is it?)
If non-competes are ever justified, they should be even less justified for convicted monopolists.