Re:Much of this could be done in linux...
on
Microsoft's new CLI
·
· Score: 1
The primary interactive use of the shell is to call exec on a list of strings. Sometimes you redirect standard io. Sometimes you also call fork. You also change the cwd. There's also a substantial amount of simple substitution from what you type to what it calls exec on.
The primary design goal of sh is that you should be able to type most strings without any characters other than the contents of the string, to optimize the common case. This conflicts with making a good scripting language, but is absolutely necessary for interactive use.
On the other hand, I think it would be useful to have a shell that was sort of like JSP, where just typing things acts like the usual shell syntax, but where you could switch to something more structured for doing control stuff. E.g.:
<% for (i in <#ls ~ #>) {<#
cp <%i%> <%i + ".bak"%> #>}%>
Where you can switch back and forth between syntaxes as appropriate for your needs.
People with the money to live in places that are unusual should spend some of that money on getting an honest assessment of the location and on measures to mitigate the danger before it happens. If you want to live on a lake, find out how high the water can get, and build higher. If you want to live in an area that burns periodically, find this out before you build the house, get yourself a buffer zone around your house, and (co-operating with local government) burn off the surrounding fuel so that it doesn't build up to these levels. The problem is not that people want or need to live in these places, it's that people think they can live in these places like they were in the middle of France or something.
But it's not like they haven't actually been making them at a plodding rate since the 80s, because I've actually used them at the MIT Media Lab. Just because they can't figure out a way to make them commercially viable doesn't mean their product doesn't exist at all.
SCO has admitted to violating the copyrights of dozens of companies and hundreds of individuals on content probably worth hundreds of millions of dollars. They are now by far the biggest pirates ever. I think it's time of all of these copyright holders to contact SCO's ISP, xo.net, and demand that SCO's site be pulled down. To do this, you send by fax or paper mail to xo.net an identification of the copyrighted work that you believe has been infringed (specifying the portions that you claim), an identification of the material that you believe to be violating it, contact information for you and for SCO, the statement "I have a good faith belief that use of the copyrighted materials described above on the allegedly infringing web pages is not authorized by the copyright owner, its agent, or the law.", the statement, "I swear, under penalty of purjury, that the information in the notification is accurate and that I am the copyright owner or am authorized to act on behalf of the owner of an exclusive right that is allegedly infringed.", and your signature. (Assuming that what google requires is what is mandated by law).
I'd enjoy the whole SCO fiasco if SCO spent the time before their case comes to trial shut off the internet for running a warez site.
I understand avoiding entity beans, because they're a major pain and don't give any benefits. And the rest of EJB isn't nearly as painful. But what benefits are you seeing from using EJB? When I used EJB, our final design had removed all of the entity beans, didn't use message-driven beans, and used session beans, but didn't actually do anything with them that couldn't be done with a simple factory pattern.
MicroOptical's stuff isn't vapor; I've actually used one of their displays, which was all nicely packaged, set up to attach to glasses, and so forth.
They seem not to be on the market yet primarily because the general market for them does not yet exist (in part because they don't sell reasonable driver hardware for the part that's at the other end of the wire from the glasses). They may also still be working on being able to mass-produce them (not too long ago, they had them available, but they were made by hand).
But if you want a HUD and you're willing to do a bit of signal processing (or get a broad fabricated, since there's a freely available design), and you can convince them to sell you one, it's available today.
Subliminal messages do work for helping you to remember something which you are already trying to remember. The classic advertizing scheme would work if the viewer was trying to remember what a bag of popcorn looks like, but if the viewer is trying to do something else (like watch a movie), it has no more effect than any other unimportant stimulus. And for the people who can't remember what popcorn looks like, you can just have a picture of it on a normal sign, and they'll remember it fine.
This is, in part, the reason that these glasses use subliminal cues: if it prompts you with some information which is irrelevant or wrong, you're not likely to be adversely affected by it. In fact, if it prompts you with a piece of information which is wrong but related to the one you're searching for, it sometimes actually helps you remember the right answer.
So subliminal messages only work in some ways, which are the ways that this application actually wants, since the point is to support your memory without either distracting you from what you're doing (imagine glasses that popped up a message every time you had to make a turn on the road; you'd read the message and drive into the car in front of you) or making you produce the wrong answer when the computer's recognition fails (which it will on non-trivial tasks).
Unfortunately, the next version of Windows will be written in Visual Haskell, which will somehow manage to have buffer overruns anyway. And all of the optimizations will be turned off because they break the non-pure GUI code.
There are two virtues to handwriting: you can do it without a computer, and you can do it in a small space. Pen and paper gives the first advantage, PDAs give the second. Tablets don't give either advantage; if you have something that large that's a computer, typing is just better. There is no reason to have a primary input method that's slow, inaccurate, large, and requires attention.
If you want to make a good tablet, make the keyboard the primary interaction, with the ability to use your finger for GUI interaction (e.g., if you're reading something on it, change pages by touching particular finger-sized spots), and the pen for drawing diagrams. Then you have a laptop that can be used to read things while closed (with the screen turned around) and can be used to make diagrams conveniently, so it's more useful than a laptop, rather than different and likely less useful.
The benefit of a hydrogen energy economy is essentially that it's the physical equivalent of an open standard. Hydrogen gas is hydrogen gas, no matter how you get it. This allows production methods to compete, rather than being tied to a particular "proprietary" standard fuel. Consider how difficult it would be to generate gasoline with a nuclear reactor. On the other hand, if you have a neighborhood power plant in a small town in Alaska, it could be used to produce hydrogen gas for the cars, and you wouldn't have to import (refined) gasoline.
Of course, this also works to deal with shortages in oil for political reasons; if the situation in the middle east is bad and oil is hard to get, you can still produce fuel (which will be more expensive).
The article is misguided on several points. The first is what you mentioned, that hydrogen threatens oil producers; oil would still be the source of the cheapest hydrogen in most circumstances.
The second is that it assumes that the persian gulf actually wants to be an oil producer, and would therefore be unhappy if oil became useless. If you were to ask people in Saudi Arabia (like some NPR reporters have), you'd find out that they don't really care too much about their oil wealth and would rather be left alone by other countries. The reason they produce and sell oil is that somebody is, in fact, literally holding a gun to their heads. The reason that oil is kept expensive is to reduce demand, not to acquire wealth.
Why not sell 2.1 instead of 3? If you buy 2.1, there's a chance you'll upgrade to 3 later. Furthermore, 2.1 has been tested and used more (since it's been around longer). If the primary source of purchases is new systems or conversions from other OSes, there's no reason to try to get people to leave the older version, unless it's hard to support.
It's pretty impressive to go 60 on 3 horsepower. (In fact, the solar car race involves batteries and the cars start out charged and sit in the sun some of the time when they aren't driving)
I'd guess (from the outputs and replacement schedules) that this plant has a core which is 1/20 as big, and the portion that gets replaced in a cycle is 1/7 as big. This means that it's a lot less difficult and dangerous to replace. (My 121 days was how long I estimated it would take to use up the amount of fuel in the 10MW reactor in a 1000MW reactor, not how long it takes to use up the fuel that's actually used in 1000MW reactors)
It's got a design where it needs mechanical energy to stay critical, so it can't break down and stay critical, and over-production won't increase the production rate. It doesn't irradiate the parts that could need to be serviced or any liquids. It contains the fuel needed for 30 years, which isn't that much in terms of a big plant (121 days supply for a normal-sized plant). Won't need to be changed for 30 years, and it'll be pretty obvious if someone tries to steal the core.
The only problem I can see with it (aside from public perception) is that it involves a shaft dug into permafrost. I'd be somewhat worried that a wet fall followed by a sudden cold spell could lead to the shaft getting crushed.
Of course, it will be hard to sell people on, despite the fact that this is probably a much safer thing to have in your back yard than a gas main. I'd like one in my back yard, except for the fact that it's not cost-effective to run, unless you're in the middle of nowhere in a place without sunlight.
By "landmarks", I mean, in the case of MapQuest, street names (and, to a certain extent, distances), which is how it tells you when you should do something. Of course, many of the places I drive don't have signs half of the time, and the streets sometimes don't really have names.
It seems to me like it should be possible to handle on ramps and off ramps by the signs and which fork to take at each branch.
I don't know about other places, but their directions in the Boston area have always been hard to follow and non-optimal. At least they fixed the bug where it had you make a turn off of an overpass onto a road you can't get on anywhere nearby.
I actually usually use MapQuest, but I generally just look at some maps and figure out the route I want by myself. The main problem with MapQuest in my experience is that it uses landmarks and turns which aren't obvious, so it helps a lot to look at the shapes of the intersections.
One of the problems that the PDA market has had for some time now is that the old ones are sufficent, so there's no need for people to get new ones. I'm still using my Visor Deluxe from years ago, because it serves my purposes (remind me to go to the dentist in 15 minutes, hold on to directions to the dentist, keep phone numbers, display e-books on occasion, and play a couple of silly games). It would be nice if it could recharge batteries in a cradle, but that's the only thing I'm really missing.
So the PDA market is largely doomed because everyone who wants one has one, and not that many people want to get new ones for the toy factor; combining with cell phones, on the other hand, reduces the number of devices people carry, so that's a worthwhile reason to buy a new device.
In any case, I expect PDAs to be used long after it becomes commercially infeasible to make a big deal about selling them. "Death" is the wrong term here; a better term might be "infertility".
The interesting point is that all of these operating systems seem to be getting faster. It seems to come down to how many recent developments have been integrated into the version being tested, not any inherent differences between operating systems. This is, of course, as it should be: the source for all of these operating systems is available, and there are even frequently papers describing the techniques. If a technique is, in fact, better, it should eventually be adopted by all of them, and so your results will depend on how much has been adopted in the version you're testing.
It is encouraging to see that all of these developers are competing with the real opponent, which is not each other or even Microsoft, but the slashdot effect. After all, the goal should not be simply to be better than the others, but to be sufficient for the user's purpose, which is not hampered but rather assisted by sharing all of your tricks. It can sometimes seem like there are endless wars between Linux and BSD, but, behind the scenes, the sides actually share information. Never as much as they'd like, but always more than people think.
I assume you mean the areas in which there is no data for FreeBSD (due to the max number of processes issue). But that's an understandable issue (although he must have missed making it runtime configurable or something). The interesting question is why FreeBSD performs better with a lot of data than with less data in some cases while still being successful.
Except that SCO's also making itself look bad (look, Linux has something similar to the code we blatantly stole from BSD and put our copywrite notice on).
Linux also doesn't depend on the market for anything particularly vital. Sure, a number of people who work on Linux have jobs that depend on the market, but a lot of people who work on Linux also have unrelated jobs. This is at most a temporary setback, and doesn't affect the possibility of Linux coming back later.
Furthermore, SCO doesn't seem to actually own any significant IP, doesn't seem to have any reasonable internal accounting, and is going up against IBM, which is a very secure position. There's practically no chance that SCO will not lose all of the pending cases, which will substantially validate the Linux position.
On the other hand, SCO has sued or is planning to sue practically everyone they've signed a contract with in the past, which the exception of Microsoft (sorry, with the *present* exception of Microsoft; as Caldera, they sued Microsoft and got a settlement). I think it's much more likely that SCO is blackmailing Microsoft. I'm sure Microsoft would pay $50 million to avoid having SCO press releases about Windows and going through discovery.
This sort of thing would be even more useful for mass transit. If people could get into their cars, say where they want to go, and have the car tell them they could get out of the car and onto the subway and get there on time, people would be much more likely to do it. As it is, it is too hard to predict mass transit times and how pleasent the trip is going to be, so people prefer to drive so that they at least feel somewhat in control of the experience.
This article doesn't even touch on that part, but I assume the idea is that there's a lot of information which is easy to prove (e.g., person X appeared on C-SPAN and said Y) and could be signed by the site collecting and inserting it. Then there's insider information which can't be verified easily, but could be signed pseudonymously, in case the source wishes to come forward in the future. Furthermore, it's sometimes worthwhile to have rumors stated explicitly, in case someone wants to come forward and confirm them.
Of course, the trick is the interface, which has to deal with different levels and types of confidence (remember that some of the information in the system will be signed by people the system tracks, who aren't any more trustworthy as sources than as officials). It's possible that the UI could make it easy to build conspiracy theories, but it will most likely produce a lot of theories that aren't what a particular theorist wants to find ("W wants cars to run on hydrogen so his oil industry stops hurting the environment so much and giving him a bad name", "Gore was conspiring with the RIAA") along with the ones the theorist wants to find ("W is a Al Queda recruiter").
Remember that one of the current scandals is over the (possible) rumor that Bush spread a (true) rumor to harm someone who confirmed a (true) rumor. In order to form a theory of any sort about this situation, you have to be as suspicious about your sources as you are about the government, since they're the same. (What if the CIA agent, upset that her husband's information wasn't getting the attention she felt it deserved, leaked her own name and started the rumor that it came from the White House? What if she wasn't even really a CIA agent, and the government's response to the leak is to cover a different person, who is an agent?)
My main server at home is a P90, which has been running nearly continuously for almost 8 years. IBM hardware can be remarkably reliable, especially considering that this machine was from a fire sale. (I've added two hard drives, some memory, and a network card, but everything that was in there originally, with the exception of an undocumented piece of plastic, is still there; it has a CD-ROM that's actually older, but I don't use it any more).
Slashdot Mirror
The primary interactive use of the shell is to call exec on a list of strings. Sometimes you redirect standard io. Sometimes you also call fork. You also change the cwd. There's also a substantial amount of simple substitution from what you type to what it calls exec on.
The primary design goal of sh is that you should be able to type most strings without any characters other than the contents of the string, to optimize the common case. This conflicts with making a good scripting language, but is absolutely necessary for interactive use.
On the other hand, I think it would be useful to have a shell that was sort of like JSP, where just typing things acts like the usual shell syntax, but where you could switch to something more structured for doing control stuff. E.g.:
<% for (i in <#ls ~ #>) {<#
cp <%i%> <%i + ".bak"%>
#>}%>
Where you can switch back and forth between syntaxes as appropriate for your needs.
People with the money to live in places that are unusual should spend some of that money on getting an honest assessment of the location and on measures to mitigate the danger before it happens. If you want to live on a lake, find out how high the water can get, and build higher. If you want to live in an area that burns periodically, find this out before you build the house, get yourself a buffer zone around your house, and (co-operating with local government) burn off the surrounding fuel so that it doesn't build up to these levels. The problem is not that people want or need to live in these places, it's that people think they can live in these places like they were in the middle of France or something.
NPR doesn't take money from individuals, but your local member station almost certainly does, and they may have a paypal link.
But it's not like they haven't actually been making them at a plodding rate since the 80s, because I've actually used them at the MIT Media Lab. Just because they can't figure out a way to make them commercially viable doesn't mean their product doesn't exist at all.
SCO has admitted to violating the copyrights of dozens of companies and hundreds of individuals on content probably worth hundreds of millions of dollars. They are now by far the biggest pirates ever. I think it's time of all of these copyright holders to contact SCO's ISP, xo.net, and demand that SCO's site be pulled down. To do this, you send by fax or paper mail to xo.net an identification of the copyrighted work that you believe has been infringed (specifying the portions that you claim), an identification of the material that you believe to be violating it, contact information for you and for SCO, the statement "I have a good faith belief that use of the copyrighted materials described above on the allegedly infringing web pages is not authorized by the copyright owner, its agent, or the law.", the statement, "I swear, under penalty of purjury, that the information in the notification is accurate and that I am the copyright owner or am authorized to act on behalf of the owner of an exclusive right that is allegedly infringed.", and your signature. (Assuming that what google requires is what is mandated by law).
I'd enjoy the whole SCO fiasco if SCO spent the time before their case comes to trial shut off the internet for running a warez site.
I understand avoiding entity beans, because they're a major pain and don't give any benefits. And the rest of EJB isn't nearly as painful. But what benefits are you seeing from using EJB? When I used EJB, our final design had removed all of the entity beans, didn't use message-driven beans, and used session beans, but didn't actually do anything with them that couldn't be done with a simple factory pattern.
MicroOptical's stuff isn't vapor; I've actually used one of their displays, which was all nicely packaged, set up to attach to glasses, and so forth.
They seem not to be on the market yet primarily because the general market for them does not yet exist (in part because they don't sell reasonable driver hardware for the part that's at the other end of the wire from the glasses). They may also still be working on being able to mass-produce them (not too long ago, they had them available, but they were made by hand).
But if you want a HUD and you're willing to do a bit of signal processing (or get a broad fabricated, since there's a freely available design), and you can convince them to sell you one, it's available today.
Subliminal messages do work for helping you to remember something which you are already trying to remember. The classic advertizing scheme would work if the viewer was trying to remember what a bag of popcorn looks like, but if the viewer is trying to do something else (like watch a movie), it has no more effect than any other unimportant stimulus. And for the people who can't remember what popcorn looks like, you can just have a picture of it on a normal sign, and they'll remember it fine.
This is, in part, the reason that these glasses use subliminal cues: if it prompts you with some information which is irrelevant or wrong, you're not likely to be adversely affected by it. In fact, if it prompts you with a piece of information which is wrong but related to the one you're searching for, it sometimes actually helps you remember the right answer.
So subliminal messages only work in some ways, which are the ways that this application actually wants, since the point is to support your memory without either distracting you from what you're doing (imagine glasses that popped up a message every time you had to make a turn on the road; you'd read the message and drive into the car in front of you) or making you produce the wrong answer when the computer's recognition fails (which it will on non-trivial tasks).
Unfortunately, the next version of Windows will be written in Visual Haskell, which will somehow manage to have buffer overruns anyway. And all of the optimizations will be turned off because they break the non-pure GUI code.
There are two virtues to handwriting: you can do it without a computer, and you can do it in a small space. Pen and paper gives the first advantage, PDAs give the second. Tablets don't give either advantage; if you have something that large that's a computer, typing is just better. There is no reason to have a primary input method that's slow, inaccurate, large, and requires attention.
If you want to make a good tablet, make the keyboard the primary interaction, with the ability to use your finger for GUI interaction (e.g., if you're reading something on it, change pages by touching particular finger-sized spots), and the pen for drawing diagrams. Then you have a laptop that can be used to read things while closed (with the screen turned around) and can be used to make diagrams conveniently, so it's more useful than a laptop, rather than different and likely less useful.
The benefit of a hydrogen energy economy is essentially that it's the physical equivalent of an open standard. Hydrogen gas is hydrogen gas, no matter how you get it. This allows production methods to compete, rather than being tied to a particular "proprietary" standard fuel. Consider how difficult it would be to generate gasoline with a nuclear reactor. On the other hand, if you have a neighborhood power plant in a small town in Alaska, it could be used to produce hydrogen gas for the cars, and you wouldn't have to import (refined) gasoline.
Of course, this also works to deal with shortages in oil for political reasons; if the situation in the middle east is bad and oil is hard to get, you can still produce fuel (which will be more expensive).
The article is misguided on several points. The first is what you mentioned, that hydrogen threatens oil producers; oil would still be the source of the cheapest hydrogen in most circumstances.
The second is that it assumes that the persian gulf actually wants to be an oil producer, and would therefore be unhappy if oil became useless. If you were to ask people in Saudi Arabia (like some NPR reporters have), you'd find out that they don't really care too much about their oil wealth and would rather be left alone by other countries.
The reason they produce and sell oil is that somebody is, in fact, literally holding a gun to their heads. The reason that oil is kept expensive is to reduce demand, not to acquire wealth.
At Microsoft Research, working on all sorts of things that the rest of Microsoft will mutilate and then put into Windows completely broken.
Why not sell 2.1 instead of 3? If you buy 2.1, there's a chance you'll upgrade to 3 later. Furthermore, 2.1 has been tested and used more (since it's been around longer). If the primary source of purchases is new systems or conversions from other OSes, there's no reason to try to get people to leave the older version, unless it's hard to support.
It's pretty impressive to go 60 on 3 horsepower. (In fact, the solar car race involves batteries and the cars start out charged and sit in the sun some of the time when they aren't driving)
I'd guess (from the outputs and replacement schedules) that this plant has a core which is 1/20 as big, and the portion that gets replaced in a cycle is 1/7 as big. This means that it's a lot less difficult and dangerous to replace. (My 121 days was how long I estimated it would take to use up the amount of fuel in the 10MW reactor in a 1000MW reactor, not how long it takes to use up the fuel that's actually used in 1000MW reactors)
It's got a design where it needs mechanical energy to stay critical, so it can't break down and stay critical, and over-production won't increase the production rate. It doesn't irradiate the parts that could need to be serviced or any liquids. It contains the fuel needed for 30 years, which isn't that much in terms of a big plant (121 days supply for a normal-sized plant). Won't need to be changed for 30 years, and it'll be pretty obvious if someone tries to steal the core.
The only problem I can see with it (aside from public perception) is that it involves a shaft dug into permafrost. I'd be somewhat worried that a wet fall followed by a sudden cold spell could lead to the shaft getting crushed.
Of course, it will be hard to sell people on, despite the fact that this is probably a much safer thing to have in your back yard than a gas main. I'd like one in my back yard, except for the fact that it's not cost-effective to run, unless you're in the middle of nowhere in a place without sunlight.
By "landmarks", I mean, in the case of MapQuest, street names (and, to a certain extent, distances), which is how it tells you when you should do something. Of course, many of the places I drive don't have signs half of the time, and the streets sometimes don't really have names.
It seems to me like it should be possible to handle on ramps and off ramps by the signs and which fork to take at each branch.
I don't know about other places, but their directions in the Boston area have always been hard to follow and non-optimal. At least they fixed the bug where it had you make a turn off of an overpass onto a road you can't get on anywhere nearby.
I actually usually use MapQuest, but I generally just look at some maps and figure out the route I want by myself. The main problem with MapQuest in my experience is that it uses landmarks and turns which aren't obvious, so it helps a lot to look at the shapes of the intersections.
One of the problems that the PDA market has had for some time now is that the old ones are sufficent, so there's no need for people to get new ones. I'm still using my Visor Deluxe from years ago, because it serves my purposes (remind me to go to the dentist in 15 minutes, hold on to directions to the dentist, keep phone numbers, display e-books on occasion, and play a couple of silly games). It would be nice if it could recharge batteries in a cradle, but that's the only thing I'm really missing.
So the PDA market is largely doomed because everyone who wants one has one, and not that many people want to get new ones for the toy factor; combining with cell phones, on the other hand, reduces the number of devices people carry, so that's a worthwhile reason to buy a new device.
In any case, I expect PDAs to be used long after it becomes commercially infeasible to make a big deal about selling them. "Death" is the wrong term here; a better term might be "infertility".
The interesting point is that all of these operating systems seem to be getting faster. It seems to come down to how many recent developments have been integrated into the version being tested, not any inherent differences between operating systems. This is, of course, as it should be: the source for all of these operating systems is available, and there are even frequently papers describing the techniques. If a technique is, in fact, better, it should eventually be adopted by all of them, and so your results will depend on how much has been adopted in the version you're testing.
It is encouraging to see that all of these developers are competing with the real opponent, which is not each other or even Microsoft, but the slashdot effect. After all, the goal should not be simply to be better than the others, but to be sufficient for the user's purpose, which is not hampered but rather assisted by sharing all of your tricks. It can sometimes seem like there are endless wars between Linux and BSD, but, behind the scenes, the sides actually share information. Never as much as they'd like, but always more than people think.
I assume you mean the areas in which there is no data for FreeBSD (due to the max number of processes issue). But that's an understandable issue (although he must have missed making it runtime configurable or something). The interesting question is why FreeBSD performs better with a lot of data than with less data in some cases while still being successful.
Except that SCO's also making itself look bad (look, Linux has something similar to the code we blatantly stole from BSD and put our copywrite notice on).
Linux also doesn't depend on the market for anything particularly vital. Sure, a number of people who work on Linux have jobs that depend on the market, but a lot of people who work on Linux also have unrelated jobs. This is at most a temporary setback, and doesn't affect the possibility of Linux coming back later.
Furthermore, SCO doesn't seem to actually own any significant IP, doesn't seem to have any reasonable internal accounting, and is going up against IBM, which is a very secure position. There's practically no chance that SCO will not lose all of the pending cases, which will substantially validate the Linux position.
On the other hand, SCO has sued or is planning to sue practically everyone they've signed a contract with in the past, which the exception of Microsoft (sorry, with the *present* exception of Microsoft; as Caldera, they sued Microsoft and got a settlement). I think it's much more likely that SCO is blackmailing Microsoft. I'm sure Microsoft would pay $50 million to avoid having SCO press releases about Windows and going through discovery.
This sort of thing would be even more useful for mass transit. If people could get into their cars, say where they want to go, and have the car tell them they could get out of the car and onto the subway and get there on time, people would be much more likely to do it. As it is, it is too hard to predict mass transit times and how pleasent the trip is going to be, so people prefer to drive so that they at least feel somewhat in control of the experience.
This article doesn't even touch on that part, but I assume the idea is that there's a lot of information which is easy to prove (e.g., person X appeared on C-SPAN and said Y) and could be signed by the site collecting and inserting it. Then there's insider information which can't be verified easily, but could be signed pseudonymously, in case the source wishes to come forward in the future. Furthermore, it's sometimes worthwhile to have rumors stated explicitly, in case someone wants to come forward and confirm them.
Of course, the trick is the interface, which has to deal with different levels and types of confidence (remember that some of the information in the system will be signed by people the system tracks, who aren't any more trustworthy as sources than as officials). It's possible that the UI could make it easy to build conspiracy theories, but it will most likely produce a lot of theories that aren't what a particular theorist wants to find ("W wants cars to run on hydrogen so his oil industry stops hurting the environment so much and giving him a bad name", "Gore was conspiring with the RIAA") along with the ones the theorist wants to find ("W is a Al Queda recruiter").
Remember that one of the current scandals is over the (possible) rumor that Bush spread a (true) rumor to harm someone who confirmed a (true) rumor. In order to form a theory of any sort about this situation, you have to be as suspicious about your sources as you are about the government, since they're the same. (What if the CIA agent, upset that her husband's information wasn't getting the attention she felt it deserved, leaked her own name and started the rumor that it came from the White House? What if she wasn't even really a CIA agent, and the government's response to the leak is to cover a different person, who is an agent?)
My main server at home is a P90, which has been running nearly continuously for almost 8 years. IBM hardware can be remarkably reliable, especially considering that this machine was from a fire sale. (I've added two hard drives, some memory, and a network card, but everything that was in there originally, with the exception of an undocumented piece of plastic, is still there; it has a CD-ROM that's actually older, but I don't use it any more).