If we've gotta do it for some reason, we must do it simply.
But the problem is that even at this level, the process is not visible and monitorable by the average voter.
And then there is the problem of a stray trace on the PC board which accidentally produces radio noise, which someone with an AM radio can monitor as he sits in a inconspicuous place watching the vote.
And after the election people who voted the wrong way have unexplained accidents happen to them.
All that dependence on cryptography adds complexities to the system that provide new potential points of corruption. Any technology which is not directly visible to and understandable by the voter provides a place to hide shenanigans.
And the issue of surveillance remains -- how do you keep an engineer from failing to seal off radio noise that can be monitored on an AM radio in the next room, where an innocuous looking person can see who is voting when, and then after the election is over people who voted the wrong way lose jobs, get important papers lost in the mail, have their children get extra attention from teachers, and worse?
VMs just provide one more place that has to be examined for strange games: the source, the compiler, the object, and, now with the VM, the interpreter.
Anytime you have the ballot see electronics close to the place and time where the voter marks and submits it, you make it possible for an engineer to accidentally design the circuit to reveal the ballot content in radio noise, thus enabling surveillance.
The unreadable ballot and the desire to read the unreadable can't be avoided. They are part of the cost of any system that can provide anonymity. We just have to put prominent notices up to remind people to look at their ballots carefully, and spoil the ballot if they are concerned that it won't be read the way they intend.
After all, it's part of the responsibility of the individual voter to see that the ballot reflects the voter's choice.
I guess I'm going to spam the thread, see if I can raise awareness on this.
Think about what an unused trace on the PC board hanging off an appropriate bit in the IO section could do. Noise can contain information. So someone accidentally designs the board to emit more radio noise than necessary, and at various precincts where people want to be able to have reprisals against people who vote the wrong way, there are innocuous looking people listening to a radio near the voting stations, where they can see who is at the voting machine at any particular time.
After the election's over and everyone's guard is down, certain people get passed over for promotions, get moved on the fast-track to "voluntary" retirement, get their insurance papers lost in the mail, or, maybe, if the low-profile stuff doesn't work, get targeted by thieves, etc.
Okay, they are thinking the right direction, but does it leave a paper trail?
Without a paper trail, you still have the question of how to check up on the local election officials.
And there is an issue that I worry about, but no one else seems to, of surveilance. Providing for re-arranging the ballot at each machine or precinct may prevent manufacturing attacks on the actual count, but there is no way to prevent the manufacturer from installing some non-obvious piece of hardware that will send apparently random radio "leakage" out on the AM band, to be decoded in the next room over where nobody is looking. Knowing who voted for whom allows reprisals, which will tend to effect the next ballot.
If netinfo tells you your user is in group wheel (or was it admin) your user is an admin user.
I had thought Apple had repented of this and was prompting to make two default users now, one admin and one working, but I guess that was just wishful thinking.
But, really, until the OS allows one to surf the web with privileges reduced below even non-admin there will be paths open for stepped escalation.
If they had not been so anxious to expand the scope of general electronic documents before the basis was in place, if they had not been so willing to adopt the internet ahead of the curve, if they had not been so willing to give us HTML in our mail,...
Okay, Apple and a bunch of others are also partly to blame. But Microsoft is the ones who pushed the competition so hard that software companies that took the time to make real products couldn't stay in business.
Microsoft has to quit selling IE and VB and even.net as solutions for business use. Either that or they set themselves up for being sued for establishing public nuisance or worse. Eventually, the banks are going to start suing MS for making false claims about their products.
The banks, also, have to start building their own special purpose browsers. A special purpose browser can force all connections to work some protocol established and encrypted by the bank, and can check both the URLs and the IP addresses against verification servers. Not impossible to game, but the speed bump can be made big enough to keep the script kiddies at bay.
There is only one way to safely access sensitive information across the net and it does not involve using a general purpose browser, and it does not involve any software from Microsoft.
This seriously smells like someone in state government trying to angle more Microsoft presence in Pennsylvania.
I mean, _public_ _money_ paying for this chance for Microsoft to experiment with kids at a far more intimate level than they already do. It's bad enough for Microsoft to mess with the kids' minds through their software.
This may be only a few hundred kids in a lower middle-class part of town, but think of the scars these kids will carry for life. Even though the kids seem to have to apply for the lotter to attend, do these kids deserve it?
the ones made by the high-paid Americans would look trashy enough anyway.
One of these days, the world will change, and people will be motivated by things other than immediate gain and loss. I don't want to advocate bad working conditions, but the worker motivated to get rewarded for putting out a good product does a better job than the worker who is guaranteed good enough conditions, whatever he thinks are good enough.
(Of course, a worker paid enough to keep self and dependents at least in good health is going to be able to focus better on doing a good job than one who is starving or worrying about starving children.)
In other words, both overpaid and underpaid workers generally do not perform as well.
And managers think that it costs less when an underpaid worker does a bad job.
might be that you assume an awful lot about the assumptions that others have.
Number one, in any particular school district, there is at least one hobbiest programer with the skills to slap together a word bingo card printing program in a reasonable small number of his spare time hours. Probably there is at least one such programer at any particular school. Okay, I'm restricting the scope of this conjecture to US schools, but I am assuming existing technologies. (I happen to have a couple such programs I have used to polish my own programing skills.)
That said, this is the sort of problem that ordinary computer users ought to be able to create such a program as a word processing document. That such is not the case reflects just how misdirected the current application tech is.
A guy with a number that low _can't_ be that clueless, can he?
Anyway, no, as others have said, once you know the box has been penetrated there is no way to be sure you've cleaned every corner where something bad can hide.
Of course, the only really safe thing to do is pull all HDs, mount them on a known clean box (preferably a different OS to provide a discontinuity), back up the data forks of the important data files, and scrub the drives with the lowest level format that the drive itself can recover from.
Unimportant data like home movies and pictures should just be written off. Hopefully, the originals are stored off-line on something not easily writable without human interaction.
On the other hand, if the user in question doesn't care whether he is unwittingly part of a botnet or potentially giving his credit card number away, by all means, just clean the malware off and keep going until it chokes up again.
Macintosh System 6 was 24 or 32 bit addressing depending on a patch and some hardware. Around 7.5, 32 bit addressing became the default (and the name was formally changed to Mac OS).
BINGO is, like, trivial to implement. Once you know your dev environment, it is literally an afternoon or two of programming.
It does make a good topic for a homework problem in an algorithms class, I suppose, or a good sample for showing the capabilities of a dev environment.
But most of the work is in understanding your dev environment, whether the old Claris/AppleWorks spreadsheet with a bunch of referenced cells and a neat range for random-sorting the referenced cells, or an old basic interpreter with its implicit mono-spaced English output and maybe line graphics, or the original Mac "system" that was more a collection of techniques than an OS, or Java (with or without netbeans or eclipse), or Apple's new project builder and the OS it's embedded in.
It would make an even better topic for homework assignment in a class on business models.
The problem here is that you want people to pay you for things they should be trivially able to do for themselves, but can't very well with the current set of tools. If you could see the tool that wants to be built here and build it, you'd have a product that people would pay you to work on, whether under the dead "proprietary" model or under the dying shareware model (which, considering how Microsoft's products get around, is the only successful "proprietary" model that ever worked), or under the more open business models of so-called f/oss.
If you can't see the tool that wants to be built here, crippleware won't take you very far either.
The numerical value in a long and the power of ten in another?
That's just floating point revisited, with 64 bits of significance (less if you mean to interpret the long as binary coded decimal).
Some floating point packages give 64 bits of significance in a long double, although it's not required by the C standard IIRC. But 64 bits only give about 18 or 19 decimal digits of significance, I forget which (but exactly 16 digits if BCD).
Changing the exponent to decimal only makes the range a bit wider. (Waters down the coverage available by not-a-numbers, too, but you probably don't mind that?)
What if the significand (your numerical value) is binary coded decimal? Then you have the advantage that all values map directly to the way we write numbers down by hand. It's a big advantage, because we no longer get surprised by things like 0.2 decimal being stored inexact. But we still have rounding errors when dividing by numbers that are multiples of any primes other than 2 and 5. (And there are lots of those.)
A way to represent repeating fraction strings (such as when you write 0.9 with a bar over the nine to show that it is the result of something like a third times 3) would help for rational numbers. But taking advantage of such a representation is not as easy as one might hope, invoking rounding issues of another sort. (What's the difference between zero point repeating nine and 1.0? How does the computer know when to resolve and how to resolve the cases that aren't as visible?) And you still leave rounding unresolved for irrational numbers such as pi and e and and multiples and fractions thereof.
(Marking the repeating end of a fraction is equivalent to keeping track of numbers as fractions, keeping both the numerator and denominator and only dividing when the result does not lose significance.)
Decimal strings have a very nice advantage, in that the visible surprises are fewer, and if you know you need 16 columns of accuracy, you can calculate the intermediates out to 32 digits of significance. But there are certain nasty surprises that tend to get swept under the rug in decimal arithmetic that binary helps resolve, so it's a kind of pick-your-poison problem, assuming you can choose between decimal strings and binary strings.
And that is really the best we can offer:
Fixed width integers for problems where we know the range doesn't exist the width of the integer.
Fixed point as a variation of fixed-width.
Fixed significance floating point for "easy" fractional problems where rounding is okay.
Arbitrary length numeric (decimal or binary) strings for the rest of the problems, remembering that we still have to deal with rounding.
Maybe sometime we'll figure out how to deal with repeating end elements, and be able to at least keep rational numbers truly accurate.
Some of the irrational numbers might be resolvable by a numbering system based on polynomial series, but I haven't seen much work on generally available libraries for those.
Take a perceived lead and pile the promises on higher and deeper, to kill whatever competition might possibly provide an alternative to Microsoft's fake technology.
Yield on the promises is always less than ten percent, and the part that I would need always gets swept under some rug and left behind.
Microsoft is the classic pusher, and Microsoft technology is intellectual crack.
Trying to remember where I found the head, perhaps it was an advertiser at lowendmac or smalldog or another of those mac special sites. I don't see it today. Well, I ignored the hits on ebay. I decided not to remember where because I think it makes more sense to get the 3rd party PS when I get tired of all the electrical tape.
One of these days I'm going to post pictures of the last repair job on my personal web site.
Slashdot Mirror
worries me.
.)
iNTEL's showcase stuff has generally been headed the wrong direction. Look at USB vs Firewire.
(As far as I am concerned, USB basically makes a mediocre replacement for the floppy disk, and that's about
And look at iNTEL's UWB and the not-invented-here attitude they've shown. They want UWB they can control the IP for, even if it doesn't do the job.
... and ordered a study from Gartner that they hoped would influence Apple to let it happen?
If we've gotta do it for some reason, we must do it simply.
But the problem is that even at this level, the process is not visible and monitorable by the average voter.
And then there is the problem of a stray trace on the PC board which accidentally produces radio noise, which someone with an AM radio can monitor as he sits in a inconspicuous place watching the vote.
And after the election people who voted the wrong way have unexplained accidents happen to them.
All that dependence on cryptography adds complexities to the system that provide new potential points of corruption. Any technology which is not directly visible to and understandable by the voter provides a place to hide shenanigans.
And the issue of surveillance remains -- how do you keep an engineer from failing to seal off radio noise that can be monitored on an AM radio in the next room, where an innocuous looking person can see who is voting when, and then after the election is over people who voted the wrong way lose jobs, get important papers lost in the mail, have their children get extra attention from teachers, and worse?
VMs just provide one more place that has to be examined for strange games: the source, the compiler, the object, and, now with the VM, the interpreter.
Anytime you have the ballot see electronics close to the place and time where the voter marks and submits it, you make it possible for an engineer to accidentally design the circuit to reveal the ballot content in radio noise, thus enabling surveillance.
The unreadable ballot and the desire to read the unreadable can't be avoided. They are part of the cost of any system that can provide anonymity. We just have to put prominent notices up to remind people to look at their ballots carefully, and spoil the ballot if they are concerned that it won't be read the way they intend.
After all, it's part of the responsibility of the individual voter to see that the ballot reflects the voter's choice.
Too easy to watch.
immediate gratification is one of the worst enemies of freedom.
I guess I'm going to spam the thread, see if I can raise awareness on this.
Think about what an unused trace on the PC board hanging off an appropriate bit in the IO section could do. Noise can contain information. So someone accidentally designs the board to emit more radio noise than necessary, and at various precincts where people want to be able to have reprisals against people who vote the wrong way, there are innocuous looking people listening to a radio near the voting stations, where they can see who is at the voting machine at any particular time.
After the election's over and everyone's guard is down, certain people get passed over for promotions, get moved on the fast-track to "voluntary" retirement, get their insurance papers lost in the mail, or, maybe, if the low-profile stuff doesn't work, get targeted by thieves, etc.
Okay, they are thinking the right direction, but does it leave a paper trail?
Without a paper trail, you still have the question of how to check up on the local election officials.
And there is an issue that I worry about, but no one else seems to, of surveilance. Providing for re-arranging the ballot at each machine or precinct may prevent manufacturing attacks on the actual count, but there is no way to prevent the manufacturer from installing some non-obvious piece of hardware that will send apparently random radio "leakage" out on the AM band, to be decoded in the next room over where nobody is looking. Knowing who voted for whom allows reprisals, which will tend to effect the next ballot.
check your user's group with netinfo.
If netinfo tells you your user is in group wheel (or was it admin) your user is an admin user.
I had thought Apple had repented of this and was prompting to make two default users now, one admin and one working, but I guess that was just wishful thinking.
But, really, until the OS allows one to surf the web with privileges reduced below even non-admin there will be paths open for stepped escalation.
Their brick and mortar probably looks about the same, and it's probably about right for the customers they are targeting.
We talk about Mom and Pop businesses on the web, well, that's what it will often look like.
seems silly to US people, I wonder if it seems so silly to Turkish people.
gauche and gawdy is right for business.
If they had not been so anxious to expand the scope of general electronic documents before the basis was in place, if they had not been so willing to adopt the internet ahead of the curve, if they had not been so willing to give us HTML in our mail, ...
.net as solutions for business use. Either that or they set themselves up for being sued for establishing public nuisance or worse. Eventually, the banks are going to start suing MS for making false claims about their products.
Okay, Apple and a bunch of others are also partly to blame. But Microsoft is the ones who pushed the competition so hard that software companies that took the time to make real products couldn't stay in business.
Microsoft has to quit selling IE and VB and even
The banks, also, have to start building their own special purpose browsers. A special purpose browser can force all connections to work some protocol established and encrypted by the bank, and can check both the URLs and the IP addresses against verification servers. Not impossible to game, but the speed bump can be made big enough to keep the script kiddies at bay.
There is only one way to safely access sensitive information across the net and it does not involve using a general purpose browser, and it does not involve any software from Microsoft.
No way.
This seriously smells like someone in state government trying to angle more Microsoft presence in Pennsylvania.
I mean, _public_ _money_ paying for this chance for Microsoft to experiment with kids at a far more intimate level than they already do. It's bad enough for Microsoft to mess with the kids' minds through their software.
This may be only a few hundred kids in a lower middle-class part of town, but think of the scars these kids will carry for life. Even though the kids seem to have to apply for the lotter to attend, do these kids deserve it?
Competency wheel?
the ones made by the high-paid Americans would look trashy enough anyway.
One of these days, the world will change, and people will be motivated by things other than immediate gain and loss. I don't want to advocate bad working conditions, but the worker motivated to get rewarded for putting out a good product does a better job than the worker who is guaranteed good enough conditions, whatever he thinks are good enough.
(Of course, a worker paid enough to keep self and dependents at least in good health is going to be able to focus better on doing a good job than one who is starving or worrying about starving children.)
In other words, both overpaid and underpaid workers generally do not perform as well.
And managers think that it costs less when an underpaid worker does a bad job.
I wonder what I am trying to say.
might be that you assume an awful lot about the assumptions that others have.
Number one, in any particular school district, there is at least one hobbiest programer with the skills to slap together a word bingo card printing program in a reasonable small number of his spare time hours. Probably there is at least one such programer at any particular school. Okay, I'm restricting the scope of this conjecture to US schools, but I am assuming existing technologies. (I happen to have a couple such programs I have used to polish my own programing skills.)
That said, this is the sort of problem that ordinary computer users ought to be able to create such a program as a word processing document. That such is not the case reflects just how misdirected the current application tech is.
A guy with a number that low _can't_ be that clueless, can he?
Anyway, no, as others have said, once you know the box has been penetrated there is no way to be sure you've cleaned every corner where something bad can hide.
Of course, the only really safe thing to do is pull all HDs, mount them on a known clean box (preferably a different OS to provide a discontinuity), back up the data forks of the important data files, and scrub the drives with the lowest level format that the drive itself can recover from.
Unimportant data like home movies and pictures should just be written off. Hopefully, the originals are stored off-line on something not easily writable without human interaction.
On the other hand, if the user in question doesn't care whether he is unwittingly part of a botnet or potentially giving his credit card number away, by all means, just clean the malware off and keep going until it chokes up again.
Macintosh System 6 was 24 or 32 bit addressing depending on a patch and some hardware. Around 7.5, 32 bit addressing became the default (and the name was formally changed to Mac OS).
BINGO is, like, trivial to implement. Once you know your dev environment, it is literally an afternoon or two of programming.
It does make a good topic for a homework problem in an algorithms class, I suppose, or a good sample for showing the capabilities of a dev environment.
But most of the work is in understanding your dev environment, whether the old Claris/AppleWorks spreadsheet with a bunch of referenced cells and a neat range for random-sorting the referenced cells, or an old basic interpreter with its implicit mono-spaced English output and maybe line graphics, or the original Mac "system" that was more a collection of techniques than an OS, or Java (with or without netbeans or eclipse), or Apple's new project builder and the OS it's embedded in.
It would make an even better topic for homework assignment in a class on business models.
The problem here is that you want people to pay you for things they should be trivially able to do for themselves, but can't very well with the current set of tools. If you could see the tool that wants to be built here and build it, you'd have a product that people would pay you to work on, whether under the dead "proprietary" model or under the dying shareware model (which, considering how Microsoft's products get around, is the only successful "proprietary" model that ever worked), or under the more open business models of so-called f/oss.
If you can't see the tool that wants to be built here, crippleware won't take you very far either.
The numerical value in a long and the power of ten in another?
That's just floating point revisited, with 64 bits of significance (less if you mean to interpret the long as binary coded decimal).
Some floating point packages give 64 bits of significance in a long double, although it's not required by the C standard IIRC. But 64 bits only give about 18 or 19 decimal digits of significance, I forget which (but exactly 16 digits if BCD).
Changing the exponent to decimal only makes the range a bit wider. (Waters down the coverage available by not-a-numbers, too, but you probably don't mind that?)
What if the significand (your numerical value) is binary coded decimal? Then you have the advantage that all values map directly to the way we write numbers down by hand. It's a big advantage, because we no longer get surprised by things like 0.2 decimal being stored inexact. But we still have rounding errors when dividing by numbers that are multiples of any primes other than 2 and 5. (And there are lots of those.)
A way to represent repeating fraction strings (such as when you write 0.9 with a bar over the nine to show that it is the result of something like a third times 3) would help for rational numbers. But taking advantage of such a representation is not as easy as one might hope, invoking rounding issues of another sort. (What's the difference between zero point repeating nine and 1.0? How does the computer know when to resolve and how to resolve the cases that aren't as visible?) And you still leave rounding unresolved for irrational numbers such as pi and e and and multiples and fractions thereof.
(Marking the repeating end of a fraction is equivalent to keeping track of numbers as fractions, keeping both the numerator and denominator and only dividing when the result does not lose significance.)
Decimal strings have a very nice advantage, in that the visible surprises are fewer, and if you know you need 16 columns of accuracy, you can calculate the intermediates out to 32 digits of significance. But there are certain nasty surprises that tend to get swept under the rug in decimal arithmetic that binary helps resolve, so it's a kind of pick-your-poison problem, assuming you can choose between decimal strings and binary strings.
And that is really the best we can offer:
Fixed width integers for problems where we know the range doesn't exist the width of the integer.
Fixed point as a variation of fixed-width.
Fixed significance floating point for "easy" fractional problems where rounding is okay.
Arbitrary length numeric (decimal or binary) strings for the rest of the problems, remembering that we still have to deal with rounding.
Maybe sometime we'll figure out how to deal with repeating end elements, and be able to at least keep rational numbers truly accurate.
Some of the irrational numbers might be resolvable by a numbering system based on polynomial series, but I haven't seen much work on generally available libraries for those.
Take a perceived lead and pile the promises on higher and deeper, to kill whatever competition might possibly provide an alternative to Microsoft's fake technology.
Yield on the promises is always less than ten percent, and the part that I would need always gets swept under some rug and left behind.
Microsoft is the classic pusher, and Microsoft technology is intellectual crack.
The apple tech note that has been posted elsewhere, at least twice under this topic (and that explains the circuitry between the shell and ground):
0 2461-en
http://developer.apple.com/qa/qa2001/qa1266.html
Oh. Lookie here: Apple's tech note to cover this whole topic:
http://docs.info.apple.com/jarticle.html?artnum=3
Trying to remember where I found the head, perhaps it was an advertiser at lowendmac or smalldog or another of those mac special sites. I don't see it today. Well, I ignored the hits on ebay. I decided not to remember where because I think it makes more sense to get the 3rd party PS when I get tired of all the electrical tape.
One of these days I'm going to post pictures of the last repair job on my personal web site.