Thanks for the reply. I did RT[F]A - my "Sheesh" was aimed at the author of the Slashdot synopsis, not the author of the article.
Hey don't blame me. Do you guys understand what it takes to get a story accepted by Slashdot? Giving an accurate and concise review of the facts gets you nowhere in this town.
Its breathless tabloidese or your post goes straight into the bit bucket. I had to spend several hours working out how to turn it into an anti-Microsoft story first.
You know, the Democrats have been gerrymandering for YEARS now, and when the shoe is finally on the other foot, they're screaming like stuck pigs. Let them stew on that a while, and see how they like it.
The problem is that DeLay has just opened up the game so that it is a continuous process. So instead of having the boundaries fixed at ten year intervals each party will commence redistricting as soon as they take over a statehouse.
I don't think that the idea of fair elections should be a partisan one. If Republicans thought about what could happen when the boot is on the other foot they might realize this as well. With the house, senate decided by tiny majorities and the whitehouse decided on a minority of the popular vote and a lawsuit the GOP can hardly claim a strong hold on power. They could easily find that they have lost everything they now have in 12 months time.
The problem has been with electoral tactics designed to shrink the electorate. Going negative as it is called. The idea of attack ads is not to get your people out to vote, its to keep the other side at home.
Also please keep in mind that a Motion to Compel is generally a rare thing. Federal and most state laws require that there be transparency between parties in a court case when it concerns evidence.
Absolutely. One other thing occurred to me. It is unusual to see a case dissmissed entirely. But it is very common, pretty much usual to see a court strike out several claims prior to trial.
This type of situation is rare precisely because it is a hopeless one. The courts do not like a plaintif who is apparently bringing a bogus case for extraneous reasons.
At the end of the SCO IBM case I predict a messy shareholder lawsuit against the officers of SCO in their personal capacity. Folk might like to notice that the Delaware courts have recently reversed their traditional supine approach to corporate governance.
Why the heck is the parent modded Troll not once, but twice!?!? If anything it should be Informative or Insightful.
The Republican party is very anxious to prevent people hearing about what they did in Florida. If they learned that the GOP was fixing the vote long before the election they might be less willing to believe the claims that the count was also rigged.
I do not know whether it was the case in this instance, but there are a couple of right wing groups that ask their members to mod down stories that they consider 'off message'.
A telling statement in the article is that in the past excesive gerrymandering would be self defeating. The concept being that the more districts you try to get in your favor the closer the win margins will be in that district,
This is the reason that the British Conservative party has all but collapsed in the UK, they have been cut to a rump after 18 years in power. Part of this is the fact that Thatcher would brook no rivals in the Cabinet and ruthlessly destroyed her most likely successors. But the bigger part is that when the Conservative controlled boundary commission reported they had gerrymandered the constituencies to give the Tories more seats by reducing their majorities and concentrating Labour support in a few constituencies. The point is that every vote over and above that needed to win the election in one constituency is in a sense wasted. Better to distribute them so that you just manage to win 6 seats in an area rather than comfortably win 4.
This meant that when their support dropped they ended up winning 2 seats or maybe none in an area they had traditionaly won 4. There is a lot of evidence that suggests that even if the Conservatives do not loose any of their current share of the vote they will lose even more seats because the UK has three parties and the progressive voters are mostly concerned with keeping the Tories out.
Would have been nice to define a not-often-used word in the article so we all don't have to dig...
The term comes from an election (in Chicago?) where the mayor (Gerry) came up with a set of fixed boundaries, one of which was in the shape of a salamander (lizard). Hence gerymander.
Any experienced pol will tall you that this type of trickery has a much bigger impact on an election than outright fixing of the polls. The way to cheat is by fixing the rules and by keeping opposing voters from the polls. During seggregation that is exactly how they stopped black people voting in Missisippi, any black man who dared to vote was liable to be lynched. The KKK and the police would man roadblocks to keep blacks from the polls and then there were the litteracy tests.
One of the big impacts on the Florida outcome was the state law that prohibits someone who has ever been convicted of a fellony from ever voting. This is another holdover from seggragation, litteracy tests were struck down but not felony disenfranchisement even though the intent (and effect) was largely the same - disproportionately disenfranchise black voters.
Click on my sig and you will see an article by a UK journalist who is one of the few who reported on this aspect of the Florida fix at the time the fix was in.
The answer BTW is not to try to fix the system to make it harder to gerrymander, change the electoral system to Single Transferable Vote and multi-member constituencies. That way you also create a way for the minor parties to be represented. With the increasing corruption of the Republican party Democrats should seriously consider this even if only as a self-interested move.
Regardless, there is a better way to get Tom DeLay and King George out of office. Get so many voters to the polls to vote against them that it does not matter how they try to rig the vote, they fail.
there was a case a few years ago where Dame Shirley Porter was convicted of ~40 million pounds worth of gerrymandering in a votes for homes scandal. Of course she's actually paid very very little of it back (less than a few hundred thousand pounds, if I remember the Private Eye story correctly)...
Thats alright, they found $30 million worth of her cash stashed away a couple of weeks ago. It is currently impounded and about to be forfeited.
The judge could, indeed, dismiss the case if SCO refuses to comply, but that is an extreme remedy that is seldom used.
The courts are seldom presented with cases as completely devoid of merit as this one.
I certainly wonder if the judge will dismiss the case out of hand if SCO turns up next to nothing yet again in 30 days time. I don't expect an outright refusal, I expect a second game of hunt the copyright violation. But I also doubt that the judge is going to allow SCO to simply slide until the next 30 days ad infinitum.
Instead I think that the court will start with remedies like a fine and sharply reduce the scope of the case, if sco will not reduce the number of claims the court will. It happens all the time in complex cases, if you claim the defendant engaged in malpractice on 10,000 occasions the judge will tell you to pick five specific instances and the case will proceed from there. If you won't choose the five instances the judge will choose for you. If you loose on those five instances you loose the whole case.
Courts are real sticklers for deadlines in cases like this where they suspect one side is manipulating the system. The appeals courts have no sympathy either. Court dockets are overcrowded enough as it is.
US, UK, France, Russia, and China have security veto power. China! HAve you considered their opinion of free speech rights?
Yeah, they lock people up without trial and deny access to a lawyer while their investigators 'interrogate' them.
Internet 'governance' comes down in the end to only two issues, who allocates IP address blocks and who allocates DNS names and under what circumstances. The issues there are simply does everyone have a right to gain access, or do some countries get to make the rules that others must follow?
The only practical issue of consequence here is does the US get to allocate an unfair proportion of IPv4 addresses now that shortages are starting to hurt? Answer - no. The second issue is does the US get to kick the Cuba domain or the Palestine domain off the net because some US politician wants to pander to a particular part of the electorate.
The answer to the last one is unfortunately a 'maybe'. If it ever happens that would be the end of US control, the root would fracture instantly and there would be a rival root run by the UN, in point of fact several of the existing roots are outside US control and could unilaterally fork. The net would behave somewhat unreliably for a while after which the UN root would be established as canonical and US influence at an end.
And of course the fact this would be the inevitable result is the reason that idiots in Congress who suggest this sort of thing get slapped down really hard. It is also the real purpose of ICANN, keep idiots in Congress out of the loop in case they push something idiotic through.
As for the Sinophobia that sweeps the US from time to time. China has a population four to five times that of the US and it is rapidly increasing industrial output. Within twenty years China and India will be the world economic superpowers. It would be better for the US to spend time thinking of how it is going to wield influence in that world than trying to isolate and daemonize.
The US has supported plenty of regimes that are considerably worse, in many cases imposing dictatorships on democracies. Ever wonder why the Iranians are so pissed with the US? Its because the US organized a coup to overthrow the democratic government and install the Shah as dictator rather than support the Iranian people as a fellow nation oppressed by a colonial tyrant.
Perhaps if the US would start by counting the votes in its own elections and keeping to the international treaties it signed it will be in a better position to lecture other countries about freedom and democracy. At the moment the stench of hypocrisy each time George Bush opens his mouth is nauseating.
Re:Soon to be decided
on
SCOrched Earth
·
· Score: 3, Informative
The magistrate has ruled. She has given SCO 30 days to reveal "with specificity" exactly what code IBM has "infringed". She also suspened all other discovery -- meaning that IBM does not have to provide SCO with anything until after SCO complies with her first order.
Suspending other discovery is amazingly significant here. It means that the judge has decided that the SCO refusal to state their claim is the central issue of the case, all other issues are now secondary.
SCO will dissemble for as long as they can manage and on about day 30 will attempt to provide a second discovery almost as broad as the first. Then we will see whether they get one last chance or a dismissal of their claims.
I don't think there will be much FUD value after a dismissal, sure SCO can appeal up to other courts and try to keep the issue live for a couple of extra months. But the stock price will be in the tank. The analysts think the price will go down to zero. The short interest in the stock is 25% of the float already and doubled in the past month.
the problem with the U.N., as contrasted to the U.S., is that the U.N. is entirely composed of states that have no equality, it is a body of unequals. No one could possibly claim that India, Russia, China and the U.S. are on par with Liberia, Syria, the Seychelles?
That is why membership of the security council is limited and the permanent members have veto power.
The story has the purpose of the summit completely wrong. There is nothing about Internet governance, it is high level touchy feely bullshit about the information society.
To find out the substance of the agenda is near impossible but the fact there are 50 heads of government there shows what is up. A meeting on controlling the Internet would be attended by ministers who do the actual dirty work, heads of government don't talk about who runs the A-Root.
The sort of thing they will be discussing is how to keep lots of languages alive in the Internet age. Popular with the International community at large but ultimately futile and we will be better off without them. Anything worth keeping will be translated into English.
People go on so about international heritage, since when has anyone mourned the fact that we no longer have a community speaking ancient Mayan or Pharonic Egyptian? What national languages are in actuality are occasion for bigottry and violence. Get rid of Basque and you get rid of Basque language nutters killing people who object to being rulled by basque language nutters.
Same goes for Welsh, one minute you have people whining about the loss of a national identity that was never really theirs in the first place, then they start imposing it on schoolkids (always a good ploy, they can't refuse and few people have the guts to stand up and object that learning Spanish or German would be a better career move), next thing welsh language loonies are burning down holiday cottages and planning lists of foreigners for 'ethinc cleansing' come the revolution. And don't get me started on the French.
Get rid of languages and you get rid of language bigotry. The Web is doing a great job in this respect. Within a couple of generations it will be impossible to hold a middle class job in any country unless you are fluent in English.
If you want your memory to run at core speed (drool...), it'll generate just as much (if not more) heat than the computational parts of the core.
No, the power consumption of CMOS is largely decided by the switching process, the leakage current is minimal. Memory runs pretty slowly because it basically sits there doing nothing. That is one reason why CPUs tend to be adding cache rather than processing capabilities. Adding memory is a lot easier.
Dynamic ram is like the name says a bit more active than static but the refresh function is independent of the clock speed.
Also, you'll need to put the memory controller on the die, too, else communications between CPU and memory would have to be routed off-die.
Yes, and all the support chips, but the need for most of those support chips goes away if you have a monolithic CPU with optical communications. This type of chip probably would NOT include a video adapter, stick that in the display device, processing is going to be much cheaper.
Re:Soon to be decided
on
SCOrched Earth
·
· Score: 2, Insightful
The IBM motion to compel filed earlier is interesting:
Specifically, IBM has alleged that SCO has, among other things, violated the Lanham Act by misrepresenting its ownership rights in Linux, tortiously interfered with IBM's prospective economic relations by making false and misleading statements to IBM's prospective customers concerning Linux, and engaged in unfair and deceptive trade practices by publishing false and disparaging statements about Linux. As it has used the specter of all its supposed rights in Linux to create uncertainty in the marketplace concerning Linux, SCO cannot deny IBM discovery of SCO's rights to Linux
IBM's motion would require SCO to reply with specificity the file and line numbers of the exact pieces of code to which they claim ownership rights and the nature of the rights claimed.
SCO refuses to do this and their excuses are pretty pathetic, they simply repeat their original accusation that IBM is guilty of misconduct. IBM is pointing out that regardless of any claim that SCO may make to prevent discovery of this information by the defense it is material to their counterclaim and that they have an absolute right to that information as a matter of law.
If the judge has any guts he will simply order SCO to reveal with specificity the exact claims made within a short period of time with a proviso that if he believes the discovery to be made in bad faith he will simply choose a set of SCO's claims at random and if IBM can show these to be baseless make a finding of fact that SCO has failled to substantiate its claim.
This is not a new game for a vexatious plaintif to play and neither is the remmedy. I doubt that the appelate courts would interfere with a district court trying to force a plaintif to provide the defense with the minimum information they require to defend themselves.
It was a reaction to the fact that extended file name sizes were available in every other OS at the time including other desktop OSs like the MacOS. Besides, you don't get a patent cookie for changing
char filename [ 12 ];
to
char filename [ 255 ];
The patent covers the specific implementation. The issue as I see it here is that under US patent law it is possible to patent a specific implementation of an interface even though the details patented are both obvious (in both the litteral and legal sense since the USPTO does not is corrupt) and non-essential.
Microsoft is forced to license this patent because it is an arbitrary interface patent. There are any number of design choices that could have been made, there is no intrinsic value in the specific choice made by Microsoft. But the USPTO will happily grant them a patent whose sole purpose is to prevent others interfacing to their system.
There are lots of examples of using the patent system to effect 'tied sales'. The pattern of slots on a razor blade handle is patented to prevent other companies offering competing blades. Lexmark and other printer companies have used the DMCA to create legal grounds to enforce an illegal tied sale.
If you look at what Microsoft is offering here they are at least offering a reasonable value in return in the form of a pretty decent compatibility testing regime. You can easily spend a couple of hundred grand on that type of testing.
Of course it does suck for OSS, but write your own device drivers and you can do whatever you like. You could even have an encrypted file store on a removable medium - somthing Microsoft seem unable to support. For some reason I can't format my compact flash cards with NTFS and enable the crypto.
That is just stupid.
That is the US Patent system, but I am repeating you
Pity that unlike 3rd world countries, getting UN observers to observe the US elections won't even help coz the machines don't have audit trails, by design.
Counting the votes is not the biggest fix in the system, choosing who gets a vote is. Back in the 1950s the southern seggregationist states had 'litteracy tests' which in practice were tests of skin color. A white guy no matter how illiterate always passed, A black guy could be a school teacher and would still be failled.
In Florida the fix was in long before the vote. Most folk know about the 'choicepoint' company that was paid to purge the voter rolls of 'fellons' and did so with an accuracy of less than 15%. Choicepoint were originally contracted to perform extensive cross checking of the scrub list but when Bush and Harris took over someone told choicepoint not to bother with any checking at all. Harris also sent numerous demands to the country returning officers on the implementation of the scrub list even though several districts were reporting astonishingly high error rates.
Click on my sig line to see the details. A postscript here is that when the Republicans had to diss the report on this fiasco they chose the same American Enterprise Institute 'fellow' to do the hatchet job who had previously been found peddling bogus statistics claiming 'More Guns Less Crime'. In fact as several researchers have shown the data Lott produced was fixed and when this spilled over into the blogosphere Lott was found to have been posting articles defending himself under a false name.
If you are worried about Diebold read the Pallast article, then work out how you can help to kick this scum out of office.
An AC wrote Shades of the GIF/Unisys fiasco. OSS FAT+, anyone?..
Thank you. GIF/Unisys is *exactly* what I was thinking of.
I seem to remember that some time ago there was a court case which involved Microsoft and anti-trust. Part of the settlement in that case was that Microsoft had to allow other companies to license its core technologies.
$250K is not an unreasonable amount for a license to what appears to be five pretty robust patents to a technology that Microsoft did in fact design. You have to make a million units to pay that fee.
If people put 5% of the effort that has gone into this flame fest into designing a new spec... Nah.
Chances of Microsoft suing the OSS community are pretty low, last patent suit I was involved in cost a couple of million in legal fees. Even though we won you don't get any of that back
Each halving of line width is a quadrupling of density, so your limit is pushed out to 32 years.
The original estimate was off since the minimum feature size is not the same as the smallest dimension in the transistor gate. The limit is actually set by the gate width which can't go less than about 5 nm without the probability of quantum tunnelling occuring going above an acceptable limit (some leakage is OK but there is a point where it is not possible to distinguish the on state from the off...
We have two more halvings of the minimum feature size before production silicon reaches state of the art and two more halvings after that before the show is all over folks.
The doubling density each year rule corresponds to a halving of minimum feature size each 2 years. Intel say this is extending to 3 years (Bush recession, war on terror etc.) and the show is over completely by about 2015 at the latest.
There will be cleanup work for some time but if you think about what you end up with at the end point it is quite interesting. You can fit 64 times the current number of transistors onto a chip. So you can comfortably fit your current state of the art Quad-Pentium processor with 4Gb of RAM all onto the same chip. So cut away the external memory bus completely and in its place add a couple of laser diodes on each edge and some receivers and you have a processing unit that communicates to its peripherals and any neighbors by means of optical couplings. It does not need special cooling either because only a small part of the chip area is CPU, the rest is memory.
Instead of adding memory to this type of system you add more processor/memory units. You cound easily fit four, sixteen into a comodity PC box. The optical couplings mean that memory paging etc can be handled by making a request to a neighboring processor that stores that information.
Big supercomputers are built by simply lashing a few hundred standard boards together.
Back to the future, I was building these systems 20 years ago. Only then we called them transputers and they only had 4Kb of RAM
T-Mobile is a flat rate of $20/mo for GPRS Internet if you have a contract with them. That's also the same price you pay for unlimited HotSpot 802.11b service per month, half-price what people who are not T-Mobile customers pay. Both kinds of access will cost you $40/mo.
They finally got a clue then. When they started they wanted MUCH more. It was utterly ridiculous. This must be a recent thing.
>break those specific machines
Confident about all of the network elements beetween those machines, then?
I don't need to be, any data can be encrypted on the network under a key that is only accessible to an application running on the right hand side (i.e. mediated access through the nexus).
What Palladium does is to shrink the boundary of the trusted computing base from the entire operating kernel to just the nexus - which Microsoft has already announced will be made available for open source review. It is a reprise of Butler Lampson's security monitor idea - not surprising since he has been heavilly involved.
There is this thing called cryptography that meets the business need you speak of.
All the controls built into Palladium are based on cryptography. It is not possible to do what Palladium does with message level cryptography. With encryption Alice can talk to Bob without disclosing to Mallet but Bob can give the document to Mallet or be careless with it allowing Mallet to get hold of it.
All the principals involved in Palladium are considered experts in the cryptography area.
Of course it doesn't do anything for me. I am locked into a two-year agreement. I can't change carriers, look into other carriers, or even think about other carriers without first being charged $170.00.
Restrictive contracts are next on the list of issues for the FCC. The number portability issue is much bigger since most contracts are for 12 months or less.
I have not heard of a 2 year contract, T-Mobile only required a one year and I got $200 bucks off the cost of the phone. Of course that would be kinda sucky if the service turned out not to work in my area. But I knew it worked in advance since other folk I know use them in my area.
The final issue will be phone portability, can you take your phone with you from one carrier to another? I don't think it will matter in the end, cell phones are little more than an pocket calculator connected to a lithium ion battery. It is the battery that is the hard to reduce cost. I don't see why phones should not be costing $10 bucks or so in a few years time without silly promotional gimics. Sure a nice model with titanium case and such may be a little more expensive, but the $350 phone will go away.
The big issue for me at this point is the ridiculous cost of data service. Data should be cheaper than voice because it is not necessary to block reserve the bandwidth. If I used any of the plans I have seen offered to date to surf the web I would be paying hundreds of dollars a week.
At the point reasonable cost data plans become available expect the number of devices with embedded phone systems to soar. I would like to have a phone in my car connected to the GPS system so it can grovel mapquest and find me routes as I am driving, also warn about trafic etc. I would also like a phone chip embedded in my laptop (along with the 802.11 card). Oh and throw away the RIM and replace it with an all function organizer.
The problem is that the data services are priced for early adopters and by now we regard the service as a commodity. Bad mix.
More accurately it means:
"People who don't trust you can trust your computer to control you."
Actually it means that people who do not trust your computer configuration can pass data to you and be confident at some level that it is not exposed.
Palladium is no better for DRM copyright enforcement applications than any other hardware technology. The problem with DRM is that it is break once run anywhere. Palladium like any other hardware enforcement system is breakable, the catch is that you have to break a system that is trusted by the sender of the data.
For copyright control you cannot be any more selective about the destination machine than requiring it to be a palladium machine. So it only taks one palladium machine ever to be broken and you are toast.
For control of sensitive company documents the issue is very different. I can configure my systems so that they only deliver sensitive data to specific palladium pcs that I have designated as trusted and to obtain my documents you have to break those specific machines.
There are still people who complain about this sort of thing. Where would the world be without corporate whistleblowers? Pretty much where we are today, there were no shortage of whistleblowers on Enron, Krugman reported repeatedly in the New York Times, few took notice until Enron collapsed and suddenly it was open season, everyone acknowledged that Enron and co had ben ripping off California...
Security is security, you can't expect technology to enforce your particular set of ethical constraints. Palladium turns out to be very useful for meeting a real business need which in most cases is completely legitimate. I do not want communications with my lawyers to be disclosed. Confidentiality is in general a good thing, it is occasionally a bad thing.
But one thing to consider is that the greater the confidence that people have that their communications are secret the greater the probability they will say something in a permenant form that later compromises them. Nixon discovered this. I don't think that security will prevent disclosure of information about criminal activities and frauds.
Take Diebold for example, if they were cluefull enough to have used DRM to control their internal documents they might have been cluefull enough to secure their Web site to stop an attacker from compromiseing their software to rig the vote. What we need in the Diebold case is not internal company memos with incriminating information. What we need is a reliable security audit.
I think the reason the US doesn't have a better train system is...
1. In the past, Ford wanted automobiles to be affordable enough so that every American family could own one.
Nah, it was decided way earlier than that and cars were not competative for long distance travel until the mid-30s. The problem with trains in the US is that the Country is simply too big and the number of train lines too small.
Trains work great in the UK (well sorta) because it is a small country sizewise with a very high population density and lots or rail lines. The quickest way of getting door to door from A to B without using a car is almost always the train.
The Japan train looks interesting for pretty much the same reason, similar population density. If the train goes at half the speed of a jet it can easily outpace a jet on time taken gate to gate on any flight less than two hours. The train may take four hours to cover the same distance but you don't have the same travel time on each end or the hour long waits to board. Train stations are almost always more convenient than airports to get to.
The place where business interests fouled up the public transit system in the US was inside the cities. One city sold its fine public transit system to a consortium of General Motors, Firestone and Standard Oil. Guess what they did with it?
Slashdot Mirror
Hey don't blame me. Do you guys understand what it takes to get a story accepted by Slashdot? Giving an accurate and concise review of the facts gets you nowhere in this town.
Its breathless tabloidese or your post goes straight into the bit bucket. I had to spend several hours working out how to turn it into an anti-Microsoft story first.
The problem is that DeLay has just opened up the game so that it is a continuous process. So instead of having the boundaries fixed at ten year intervals each party will commence redistricting as soon as they take over a statehouse.
I don't think that the idea of fair elections should be a partisan one. If Republicans thought about what could happen when the boot is on the other foot they might realize this as well. With the house, senate decided by tiny majorities and the whitehouse decided on a minority of the popular vote and a lawsuit the GOP can hardly claim a strong hold on power. They could easily find that they have lost everything they now have in 12 months time.
The problem has been with electoral tactics designed to shrink the electorate. Going negative as it is called. The idea of attack ads is not to get your people out to vote, its to keep the other side at home.
Absolutely. One other thing occurred to me. It is unusual to see a case dissmissed entirely. But it is very common, pretty much usual to see a court strike out several claims prior to trial.
This type of situation is rare precisely because it is a hopeless one. The courts do not like a plaintif who is apparently bringing a bogus case for extraneous reasons.
At the end of the SCO IBM case I predict a messy shareholder lawsuit against the officers of SCO in their personal capacity. Folk might like to notice that the Delaware courts have recently reversed their traditional supine approach to corporate governance.
The Republican party is very anxious to prevent people hearing about what they did in Florida. If they learned that the GOP was fixing the vote long before the election they might be less willing to believe the claims that the count was also rigged.
I do not know whether it was the case in this instance, but there are a couple of right wing groups that ask their members to mod down stories that they consider 'off message'.
This is the reason that the British Conservative party has all but collapsed in the UK, they have been cut to a rump after 18 years in power. Part of this is the fact that Thatcher would brook no rivals in the Cabinet and ruthlessly destroyed her most likely successors. But the bigger part is that when the Conservative controlled boundary commission reported they had gerrymandered the constituencies to give the Tories more seats by reducing their majorities and concentrating Labour support in a few constituencies. The point is that every vote over and above that needed to win the election in one constituency is in a sense wasted. Better to distribute them so that you just manage to win 6 seats in an area rather than comfortably win 4.
This meant that when their support dropped they ended up winning 2 seats or maybe none in an area they had traditionaly won 4. There is a lot of evidence that suggests that even if the Conservatives do not loose any of their current share of the vote they will lose even more seats because the UK has three parties and the progressive voters are mostly concerned with keeping the Tories out.
The term comes from an election (in Chicago?) where the mayor (Gerry) came up with a set of fixed boundaries, one of which was in the shape of a salamander (lizard). Hence gerymander.
Any experienced pol will tall you that this type of trickery has a much bigger impact on an election than outright fixing of the polls. The way to cheat is by fixing the rules and by keeping opposing voters from the polls. During seggregation that is exactly how they stopped black people voting in Missisippi, any black man who dared to vote was liable to be lynched. The KKK and the police would man roadblocks to keep blacks from the polls and then there were the litteracy tests.
One of the big impacts on the Florida outcome was the state law that prohibits someone who has ever been convicted of a fellony from ever voting. This is another holdover from seggragation, litteracy tests were struck down but not felony disenfranchisement even though the intent (and effect) was largely the same - disproportionately disenfranchise black voters.
Click on my sig and you will see an article by a UK journalist who is one of the few who reported on this aspect of the Florida fix at the time the fix was in.
The answer BTW is not to try to fix the system to make it harder to gerrymander, change the electoral system to Single Transferable Vote and multi-member constituencies. That way you also create a way for the minor parties to be represented. With the increasing corruption of the Republican party Democrats should seriously consider this even if only as a self-interested move.
Regardless, there is a better way to get Tom DeLay and King George out of office. Get so many voters to the polls to vote against them that it does not matter how they try to rig the vote, they fail.
Thats alright, they found $30 million worth of her cash stashed away a couple of weeks ago. It is currently impounded and about to be forfeited.
The courts are seldom presented with cases as completely devoid of merit as this one.
I certainly wonder if the judge will dismiss the case out of hand if SCO turns up next to nothing yet again in 30 days time. I don't expect an outright refusal, I expect a second game of hunt the copyright violation. But I also doubt that the judge is going to allow SCO to simply slide until the next 30 days ad infinitum.
Instead I think that the court will start with remedies like a fine and sharply reduce the scope of the case, if sco will not reduce the number of claims the court will. It happens all the time in complex cases, if you claim the defendant engaged in malpractice on 10,000 occasions the judge will tell you to pick five specific instances and the case will proceed from there. If you won't choose the five instances the judge will choose for you. If you loose on those five instances you loose the whole case.
Courts are real sticklers for deadlines in cases like this where they suspect one side is manipulating the system. The appeals courts have no sympathy either. Court dockets are overcrowded enough as it is.
Yeah, they lock people up without trial and deny access to a lawyer while their investigators 'interrogate' them.
Internet 'governance' comes down in the end to only two issues, who allocates IP address blocks and who allocates DNS names and under what circumstances. The issues there are simply does everyone have a right to gain access, or do some countries get to make the rules that others must follow?
The only practical issue of consequence here is does the US get to allocate an unfair proportion of IPv4 addresses now that shortages are starting to hurt? Answer - no. The second issue is does the US get to kick the Cuba domain or the Palestine domain off the net because some US politician wants to pander to a particular part of the electorate.
The answer to the last one is unfortunately a 'maybe'. If it ever happens that would be the end of US control, the root would fracture instantly and there would be a rival root run by the UN, in point of fact several of the existing roots are outside US control and could unilaterally fork. The net would behave somewhat unreliably for a while after which the UN root would be established as canonical and US influence at an end.
And of course the fact this would be the inevitable result is the reason that idiots in Congress who suggest this sort of thing get slapped down really hard. It is also the real purpose of ICANN, keep idiots in Congress out of the loop in case they push something idiotic through.
As for the Sinophobia that sweeps the US from time to time. China has a population four to five times that of the US and it is rapidly increasing industrial output. Within twenty years China and India will be the world economic superpowers. It would be better for the US to spend time thinking of how it is going to wield influence in that world than trying to isolate and daemonize.
The US has supported plenty of regimes that are considerably worse, in many cases imposing dictatorships on democracies. Ever wonder why the Iranians are so pissed with the US? Its because the US organized a coup to overthrow the democratic government and install the Shah as dictator rather than support the Iranian people as a fellow nation oppressed by a colonial tyrant.
Perhaps if the US would start by counting the votes in its own elections and keeping to the international treaties it signed it will be in a better position to lecture other countries about freedom and democracy. At the moment the stench of hypocrisy each time George Bush opens his mouth is nauseating.
Suspending other discovery is amazingly significant here. It means that the judge has decided that the SCO refusal to state their claim is the central issue of the case, all other issues are now secondary.
SCO will dissemble for as long as they can manage and on about day 30 will attempt to provide a second discovery almost as broad as the first. Then we will see whether they get one last chance or a dismissal of their claims.
I don't think there will be much FUD value after a dismissal, sure SCO can appeal up to other courts and try to keep the issue live for a couple of extra months. But the stock price will be in the tank. The analysts think the price will go down to zero. The short interest in the stock is 25% of the float already and doubled in the past month.
That is why membership of the security council is limited and the permanent members have veto power.
The story has the purpose of the summit completely wrong. There is nothing about Internet governance, it is high level touchy feely bullshit about the information society.
To find out the substance of the agenda is near impossible but the fact there are 50 heads of government there shows what is up. A meeting on controlling the Internet would be attended by ministers who do the actual dirty work, heads of government don't talk about who runs the A-Root.
The sort of thing they will be discussing is how to keep lots of languages alive in the Internet age. Popular with the International community at large but ultimately futile and we will be better off without them. Anything worth keeping will be translated into English.
People go on so about international heritage, since when has anyone mourned the fact that we no longer have a community speaking ancient Mayan or Pharonic Egyptian? What national languages are in actuality are occasion for bigottry and violence. Get rid of Basque and you get rid of Basque language nutters killing people who object to being rulled by basque language nutters.
Same goes for Welsh, one minute you have people whining about the loss of a national identity that was never really theirs in the first place, then they start imposing it on schoolkids (always a good ploy, they can't refuse and few people have the guts to stand up and object that learning Spanish or German would be a better career move), next thing welsh language loonies are burning down holiday cottages and planning lists of foreigners for 'ethinc cleansing' come the revolution. And don't get me started on the French.
Get rid of languages and you get rid of language bigotry. The Web is doing a great job in this respect. Within a couple of generations it will be impossible to hold a middle class job in any country unless you are fluent in English.
No, the power consumption of CMOS is largely decided by the switching process, the leakage current is minimal. Memory runs pretty slowly because it basically sits there doing nothing. That is one reason why CPUs tend to be adding cache rather than processing capabilities. Adding memory is a lot easier.
Dynamic ram is like the name says a bit more active than static but the refresh function is independent of the clock speed.
Also, you'll need to put the memory controller on the die, too, else communications between CPU and memory would have to be routed off-die.
Yes, and all the support chips, but the need for most of those support chips goes away if you have a monolithic CPU with optical communications. This type of chip probably would NOT include a video adapter, stick that in the display device, processing is going to be much cheaper.
IBM's motion would require SCO to reply with specificity the file and line numbers of the exact pieces of code to which they claim ownership rights and the nature of the rights claimed.
SCO refuses to do this and their excuses are pretty pathetic, they simply repeat their original accusation that IBM is guilty of misconduct. IBM is pointing out that regardless of any claim that SCO may make to prevent discovery of this information by the defense it is material to their counterclaim and that they have an absolute right to that information as a matter of law.
If the judge has any guts he will simply order SCO to reveal with specificity the exact claims made within a short period of time with a proviso that if he believes the discovery to be made in bad faith he will simply choose a set of SCO's claims at random and if IBM can show these to be baseless make a finding of fact that SCO has failled to substantiate its claim.
This is not a new game for a vexatious plaintif to play and neither is the remmedy. I doubt that the appelate courts would interfere with a district court trying to force a plaintif to provide the defense with the minimum information they require to defend themselves.
The patent covers the specific implementation. The issue as I see it here is that under US patent law it is possible to patent a specific implementation of an interface even though the details patented are both obvious (in both the litteral and legal sense since the USPTO does not is corrupt) and non-essential.
Microsoft is forced to license this patent because it is an arbitrary interface patent. There are any number of design choices that could have been made, there is no intrinsic value in the specific choice made by Microsoft. But the USPTO will happily grant them a patent whose sole purpose is to prevent others interfacing to their system.
There are lots of examples of using the patent system to effect 'tied sales'. The pattern of slots on a razor blade handle is patented to prevent other companies offering competing blades. Lexmark and other printer companies have used the DMCA to create legal grounds to enforce an illegal tied sale.
If you look at what Microsoft is offering here they are at least offering a reasonable value in return in the form of a pretty decent compatibility testing regime. You can easily spend a couple of hundred grand on that type of testing.
Of course it does suck for OSS, but write your own device drivers and you can do whatever you like. You could even have an encrypted file store on a removable medium - somthing Microsoft seem unable to support. For some reason I can't format my compact flash cards with NTFS and enable the crypto.
That is just stupid.
That is the US Patent system, but I am repeating you
The patents relate to the mechanisms used to implement extended file names.
Counting the votes is not the biggest fix in the system, choosing who gets a vote is. Back in the 1950s the southern seggregationist states had 'litteracy tests' which in practice were tests of skin color. A white guy no matter how illiterate always passed, A black guy could be a school teacher and would still be failled.
In Florida the fix was in long before the vote. Most folk know about the 'choicepoint' company that was paid to purge the voter rolls of 'fellons' and did so with an accuracy of less than 15%. Choicepoint were originally contracted to perform extensive cross checking of the scrub list but when Bush and Harris took over someone told choicepoint not to bother with any checking at all. Harris also sent numerous demands to the country returning officers on the implementation of the scrub list even though several districts were reporting astonishingly high error rates.
Click on my sig line to see the details. A postscript here is that when the Republicans had to diss the report on this fiasco they chose the same American Enterprise Institute 'fellow' to do the hatchet job who had previously been found peddling bogus statistics claiming 'More Guns Less Crime'. In fact as several researchers have shown the data Lott produced was fixed and when this spilled over into the blogosphere Lott was found to have been posting articles defending himself under a false name.
If you are worried about Diebold read the Pallast article, then work out how you can help to kick this scum out of office.
I seem to remember that some time ago there was a court case which involved Microsoft and anti-trust. Part of the settlement in that case was that Microsoft had to allow other companies to license its core technologies.
$250K is not an unreasonable amount for a license to what appears to be five pretty robust patents to a technology that Microsoft did in fact design. You have to make a million units to pay that fee.
If people put 5% of the effort that has gone into this flame fest into designing a new spec... Nah.
Chances of Microsoft suing the OSS community are pretty low, last patent suit I was involved in cost a couple of million in legal fees. Even though we won you don't get any of that back
I guess someone did not read their 1040 too closely for that year. The $300 'refund' was clawed back. It was a gimmick.
Over the next thirty years the interest we will each pay on the Bush deficit will come to rather more than $300 a year.
Pannem et circenses.
The original estimate was off since the minimum feature size is not the same as the smallest dimension in the transistor gate. The limit is actually set by the gate width which can't go less than about 5 nm without the probability of quantum tunnelling occuring going above an acceptable limit (some leakage is OK but there is a point where it is not possible to distinguish the on state from the off...
We have two more halvings of the minimum feature size before production silicon reaches state of the art and two more halvings after that before the show is all over folks.
The doubling density each year rule corresponds to a halving of minimum feature size each 2 years. Intel say this is extending to 3 years (Bush recession, war on terror etc.) and the show is over completely by about 2015 at the latest.
There will be cleanup work for some time but if you think about what you end up with at the end point it is quite interesting. You can fit 64 times the current number of transistors onto a chip. So you can comfortably fit your current state of the art Quad-Pentium processor with 4Gb of RAM all onto the same chip. So cut away the external memory bus completely and in its place add a couple of laser diodes on each edge and some receivers and you have a processing unit that communicates to its peripherals and any neighbors by means of optical couplings. It does not need special cooling either because only a small part of the chip area is CPU, the rest is memory.
Instead of adding memory to this type of system you add more processor/memory units. You cound easily fit four, sixteen into a comodity PC box. The optical couplings mean that memory paging etc can be handled by making a request to a neighboring processor that stores that information.
Big supercomputers are built by simply lashing a few hundred standard boards together.
Back to the future, I was building these systems 20 years ago. Only then we called them transputers and they only had 4Kb of RAM
They finally got a clue then. When they started they wanted MUCH more. It was utterly ridiculous. This must be a recent thing.
I might even get a Wireless Pocket PC...
Confident about all of the network elements beetween those machines, then?
I don't need to be, any data can be encrypted on the network under a key that is only accessible to an application running on the right hand side (i.e. mediated access through the nexus).
What Palladium does is to shrink the boundary of the trusted computing base from the entire operating kernel to just the nexus - which Microsoft has already announced will be made available for open source review. It is a reprise of Butler Lampson's security monitor idea - not surprising since he has been heavilly involved.
All the controls built into Palladium are based on cryptography. It is not possible to do what Palladium does with message level cryptography. With encryption Alice can talk to Bob without disclosing to Mallet but Bob can give the document to Mallet or be careless with it allowing Mallet to get hold of it.
All the principals involved in Palladium are considered experts in the cryptography area.
Restrictive contracts are next on the list of issues for the FCC. The number portability issue is much bigger since most contracts are for 12 months or less.
I have not heard of a 2 year contract, T-Mobile only required a one year and I got $200 bucks off the cost of the phone. Of course that would be kinda sucky if the service turned out not to work in my area. But I knew it worked in advance since other folk I know use them in my area.
The final issue will be phone portability, can you take your phone with you from one carrier to another? I don't think it will matter in the end, cell phones are little more than an pocket calculator connected to a lithium ion battery. It is the battery that is the hard to reduce cost. I don't see why phones should not be costing $10 bucks or so in a few years time without silly promotional gimics. Sure a nice model with titanium case and such may be a little more expensive, but the $350 phone will go away.
The big issue for me at this point is the ridiculous cost of data service. Data should be cheaper than voice because it is not necessary to block reserve the bandwidth. If I used any of the plans I have seen offered to date to surf the web I would be paying hundreds of dollars a week.
At the point reasonable cost data plans become available expect the number of devices with embedded phone systems to soar. I would like to have a phone in my car connected to the GPS system so it can grovel mapquest and find me routes as I am driving, also warn about trafic etc. I would also like a phone chip embedded in my laptop (along with the 802.11 card). Oh and throw away the RIM and replace it with an all function organizer.
The problem is that the data services are priced for early adopters and by now we regard the service as a commodity. Bad mix.
Actually it means that people who do not trust your computer configuration can pass data to you and be confident at some level that it is not exposed.
Palladium is no better for DRM copyright enforcement applications than any other hardware technology. The problem with DRM is that it is break once run anywhere. Palladium like any other hardware enforcement system is breakable, the catch is that you have to break a system that is trusted by the sender of the data.
For copyright control you cannot be any more selective about the destination machine than requiring it to be a palladium machine. So it only taks one palladium machine ever to be broken and you are toast.
For control of sensitive company documents the issue is very different. I can configure my systems so that they only deliver sensitive data to specific palladium pcs that I have designated as trusted and to obtain my documents you have to break those specific machines.
There are still people who complain about this sort of thing. Where would the world be without corporate whistleblowers? Pretty much where we are today, there were no shortage of whistleblowers on Enron, Krugman reported repeatedly in the New York Times, few took notice until Enron collapsed and suddenly it was open season, everyone acknowledged that Enron and co had ben ripping off California...
Security is security, you can't expect technology to enforce your particular set of ethical constraints. Palladium turns out to be very useful for meeting a real business need which in most cases is completely legitimate. I do not want communications with my lawyers to be disclosed. Confidentiality is in general a good thing, it is occasionally a bad thing.
But one thing to consider is that the greater the confidence that people have that their communications are secret the greater the probability they will say something in a permenant form that later compromises them. Nixon discovered this. I don't think that security will prevent disclosure of information about criminal activities and frauds.
Take Diebold for example, if they were cluefull enough to have used DRM to control their internal documents they might have been cluefull enough to secure their Web site to stop an attacker from compromiseing their software to rig the vote. What we need in the Diebold case is not internal company memos with incriminating information. What we need is a reliable security audit.
Nah, it was decided way earlier than that and cars were not competative for long distance travel until the mid-30s. The problem with trains in the US is that the Country is simply too big and the number of train lines too small.
Trains work great in the UK (well sorta) because it is a small country sizewise with a very high population density and lots or rail lines. The quickest way of getting door to door from A to B without using a car is almost always the train.
The Japan train looks interesting for pretty much the same reason, similar population density. If the train goes at half the speed of a jet it can easily outpace a jet on time taken gate to gate on any flight less than two hours. The train may take four hours to cover the same distance but you don't have the same travel time on each end or the hour long waits to board. Train stations are almost always more convenient than airports to get to.
The place where business interests fouled up the public transit system in the US was inside the cities. One city sold its fine public transit system to a consortium of General Motors, Firestone and Standard Oil. Guess what they did with it?