We have been concentrating too much on the post-election auditing, and not enough on the pre-election requirements, design, build test cycle.
A short proposal:
A) A vendor is chosen (Not Diebold, because they obviously cannot do it)
B) Vendor designs the system, chooses the tools, and builds it.
C) 3 validation teams are chosen. 1 commercial entity, 1 university entity, and one independant team chosen by 2 or 3 of the major parties. Maybe even us, the general public as well.
D) The teams audit and validate the code, processes, and fully test the system
E) At this point...the entire system is locked away. The vendor no longer has access to it. No patches, no updates, no more code.
E.1) The system needs to be treated with no less security than a nuclear weapon in the military. i.e. it becomes a "No Lone Zone" - No one person or entity has acceess to the system. There MUST be a team of at least two distinct parties with the keys to the storage warehouse.
F) On election day, the parties with the keys unlock their respective locks, and the system is put in place.
G) Vote
H) 3 other disparate but similar teams, validate the results. Paper against the actual db.
I) No Profit, but a valid election.
Steps C-E.1 are, currently, completely, ignored. That's what needs to change.
4) Card is then re-inserted into machine to generate the backup tally.
If the tallies from 1 and 4 don't match, the cards are "certified" and then rerun.
And the software has a routine in it to report a 'match' between the real and the backup tally, no matter what the actual result.
At any point in the stream, the s/w could be made to report whatever the people who wrote want. Unless there is a LOT more stringent pre-election auditing and security.
All of those checks and balances things need to be in place. But first, the thing needs to be written and verified correctly. It appears that Diebold did a staggeringly bad job at this.
25% error rate Patches applied without testing or verification Operating system crashing
Reliable, accurate database systems are not that hard. Companies have been doing them for decades.
Why can't Diebold do it right? Because no one held their feet to the fire to ensure it got done right. A company has employees and or stockholders to help insure that their results are correct. Diebold, in this case, had no one.
If a db application I had written had a 25% error rate, at the very least I'd be out of a job. Maybe even the company I work for would go down the tubes.
Maybe that's what should happen with Diebold. Fire them and get someone in who can actually do a valid job.
Except that one side really is rigging elections and the other isn't. Kinda like wishing to kill someone vs hiring a hitman.
And democrats (or persons wanting to getting a democrat elected) have never rigged an election? Please. The only thing the repub's did worse in this instance, was doing it badly. So badly, that they got (are getting) caught.
Just have the machine produce a paper ballot, the voter can then verify it visually, and the ballot is dropped in a lock box at the polling location.
That still has no bearing on what the db records. The person selects X, the paper prints X, the db records Y. The db recor is the one that counts.
Only in the case of a demanded recount would the paper come out. I can think of a few scenarios where it would be close, and altered, but not close enough to demand a recount.
Swing a few votes for the 3rd place guy to "your" guy. No one notices that the Loon Party has lost a few percentage points, and you gain some. Might only have to happen in a small number of districts to affect the overall outcome.
The voting machines run on version 3.0 of Windows CE, he said, and they patched it to correct problems they were having with the system.
Still, that's a problem with Diebold, not necessarily MS. Requirement definition prior to choosing the tool to do the job. C'mon...how many times have we looked at a requirement, and had to check out several different tools which may or may not do the job? And then had to throw out most, or all but one, when implementing it?
If CE was not up to the task when they started writing code for it....why use it?
Evidently, they waited until just prior to the election to find out that it wasn't really working.
Sounds like someone chose a 'solution' before doing a real analysis of the requirement.
By the looks of it, this could have been written with Linux and mysql, and still had the same problems. These guys do not know WTF they are doing.
Cowboy programming, no matter what the tool, is always a bad idea.
And a corrupt Democrat in that situation would have done exactly the same thing. Neither side is better or worse than the other.
It is evidently up to us to keep those fools straight. Demand code and process verification prior to the election. Run test after test, in public. Audit the test results. Lock the code and machines away, by a 3rd party, until actual use. Audit the results.
Still doesn't work. If you produce a paper copy, either for the voter, or to store away...then you have two different 'votes', cast by the same person. One on paper, and one in the db.
The paper says one thing, the database record may well say something else. And you, the voter, would never know. And what do you do in case of a close election? Have everyone bring back their hardcopies and count those?
Taking home a hardcopy then brings the situation of linking a voter to a vote. "If you expect to get paid this week, bring me your voting receipt that says you voted for Mr. X."
An aircraft launched missile has the advantage of being where you want it, when you want it.
Assuming the target is in low enough orbit, and fighter jet launching a missile can hit it in any orbit. A land based laser system has to wait until it is overhead.
The F-15 based ASAT system was not merely a proposal, but a developed and launch tested system.
Acquire the target, zoom up to about 80,000, launch. The missile takes over from there.
Within a few hours, you can hit a LEO sat in any orbital inclination around the globe.
It's what we want, and it's what they want. (They being the entertainment industry, the media, whatever you want to call 'em)
However...the two parties want different types of control.
They want a return to before the 60's. When they had sole reign over what you saw, when you saw it, and how you paid for it. Want to watch a TV show? Have to wait until they air it, and watch the commercials they want you to. Movies? Have to pay for each and every viewing. On their schedule.
Lately, the scale has tipped more towards us. We can (sort of) watch what we want, when and how we want to. A TV show? Sure...tape it, and watch it later. Or next year. But we've only 'paid' for it (by viewing the ads) once, if at all. And they've only goten paid once by the ad agency. Movies? Get a pirate copy, and watch it over and over. Give it to your friends. Hell...give it to everyone.
They want that control back. They would like us to (again) pay every time we 'enjoy' some of their fine offerings. Pay either directly in money, or by watching ads.
Ultimately, I think they'd like for all the hard copy ability to go away. Pure streaming. Then, they can change things, insert new ads, get us to pay over and over and over again, for the same basic content.
If you watch a rerun on TV 2 years later, of course they don't have the same ads in it. They've gotten their consumers to pay twice for the same content. And they'd like that to continue. With a VCR, PVR, whatever...a 'rerun' in your home will have the same ads in it as did the original. They've only gotten paid once.
The question is...who gets to decide who the control will reside with? CBS, et al, would obviously like to regain control over what, when, where, and how thay get paid. Currently, the control is (somewhat) in our hands. And getting more so everyday. Lets try to keep it that way.
Remember...we are not the consumers. We are the product. The ad agencies are the consumers.
Really? Look at a contrail some time after the aircraft goes by. It is quite diffuse.
There could have been localized winds, stirring up that part faster then the rest of it. Leading to a wider appearing cloud of ice particles. Which then reflect the setting sun.
Others have said..."That's just not possible! I need to drive."
How about something in the middle?
Want to reduce rush hour traffic by an easy 10%? Find an alternate way to work twice a month. That's pretty much all it would take. Not everyday. Not even every other day. If we could average alternate transport (carpool, bus, bike, whatever) twice a month, the problems would go down significantly.
I'm doing enough for about 5 of you (ride my bike 2-3 times a week).
Because the EULA is not for your benefit, but for theirs. If it were in plain English, the average user might actually be cognizant of the actual restrictions of what he is agreeing to, and then not do it.
Couched in legalese, you probably won't notice that the average EULA gives you, the consumer, no rights whatsoever, and protects the vendor from any action on your part.
$.25 is not unreasonable. It's the particulars of this plan that does kind of suck. You have a very low, hard limit of how many tracks you can d/l. And you're being charged that whether you use it or not.
There is no carryover of unused tracks.
Currently, there is no way to actually buy more. Either 40 tracks for $10, or 65 for $15. Unless you can and want to buy into the $50 plan.
emusic has a very eclectic catalog. Part of the draw there was being able to sample different artists at will. Now...you'd have to pick and choose very carefully what you wanted to spend your 40 tracks on.
It's 'bad' because now you get less for the same money.
I've been with them almost a year. My sub runs out in Nov. (Now...it's Nov 7 to be exact)
In that time, I've grabbed about 130 cd's. So maybe 12 cd's per month. 120 tracks on average. Often, I might go a month or two without anything, and then go get a bunch all at once.
With this new d/l limit, I'd have to cut back to 1/3. About 4 cd's per month, for the same price. And no month to month carryover of unused tracks.
Plus, now you'd have to be MUCH more careful about which tracks you actually d/l. Gone will be the concept of "just get the whole album". If I were to continue, I'd pick and choose each track so as to maximise my selections. Previous, if a few tracks on the album sucked...so what. It didn't cost anything extra.
but they are company trying to make a profit.
Right. My question is...were they making a profit before, or is it simply a case of the new owners wanting to make more profit? IMO, they are making a big mistake, and there will be a mass exodus fo current subscribers.
Slashdot Mirror
We have been concentrating too much on the post-election auditing, and not enough on the pre-election requirements, design, build test cycle.
A short proposal:
A) A vendor is chosen (Not Diebold, because they obviously cannot do it)
B) Vendor designs the system, chooses the tools, and builds it.
C) 3 validation teams are chosen. 1 commercial entity, 1 university entity, and one independant team chosen by 2 or 3 of the major parties. Maybe even us, the general public as well.
D) The teams audit and validate the code, processes, and fully test the system
E) At this point...the entire system is locked away. The vendor no longer has access to it. No patches, no updates, no more code.
E.1) The system needs to be treated with no less security than a nuclear weapon in the military. i.e. it becomes a "No Lone Zone" - No one person or entity has acceess to the system. There MUST be a team of at least two distinct parties with the keys to the storage warehouse.
F) On election day, the parties with the keys unlock their respective locks, and the system is put in place.
G) Vote
H) 3 other disparate but similar teams, validate the results. Paper against the actual db.
I) No Profit, but a valid election.
Steps C-E.1 are, currently, completely, ignored. That's what needs to change.
4) Card is then re-inserted into machine to generate the backup tally.
If the tallies from 1 and 4 don't match, the cards are "certified" and then rerun.
And the software has a routine in it to report a 'match' between the real and the backup tally, no matter what the actual result.
At any point in the stream, the s/w could be made to report whatever the people who wrote want.
Unless there is a LOT more stringent pre-election auditing and security.
All of those checks and balances things need to be in place. But first, the thing needs to be written and verified correctly. It appears that Diebold did a staggeringly bad job at this.
25% error rate
Patches applied without testing or verification
Operating system crashing
Reliable, accurate database systems are not that hard. Companies have been doing them for decades.
Why can't Diebold do it right? Because no one held their feet to the fire to ensure it got done right. A company has employees and or stockholders to help insure that their results are correct. Diebold, in this case, had no one.
If a db application I had written had a 25% error rate, at the very least I'd be out of a job. Maybe even the company I work for would go down the tubes.
Maybe that's what should happen with Diebold. Fire them and get someone in who can actually do a valid job.
Except that one side really is rigging elections and the other isn't. Kinda like wishing to kill someone vs hiring a hitman.
And democrats (or persons wanting to getting a democrat elected) have never rigged an election? Please.
The only thing the repub's did worse in this instance, was doing it badly. So badly, that they got (are getting) caught.
Just have the machine produce a paper ballot, the voter can then verify it visually, and the ballot is dropped in a lock box at the polling location.
That still has no bearing on what the db records. The person selects X, the paper prints X, the db records Y. The db recor is the one that counts.
Only in the case of a demanded recount would the paper come out. I can think of a few scenarios where it would be close, and altered, but not close enough to demand a recount.
Swing a few votes for the 3rd place guy to "your" guy. No one notices that the Loon Party has lost a few percentage points, and you gain some. Might only have to happen in a small number of districts to affect the overall outcome.
Right. Tested, a few missiles produced, and then cancelled.
...missiles. Not that special. No reason why a few more couldn't be built in time of need.
The launch aircraft still exist, and the missiles are
The voting machines run on version 3.0 of Windows CE, he said, and they patched it to correct problems they were having with the system.
Still, that's a problem with Diebold, not necessarily MS.
Requirement definition prior to choosing the tool to do the job. C'mon...how many times have we looked at a requirement, and had to check out several different tools which may or may not do the job? And then had to throw out most, or all but one, when implementing it?
If CE was not up to the task when they started writing code for it....why use it?
Evidently, they waited until just prior to the election to find out that it wasn't really working.
Sounds like someone chose a 'solution' before doing a real analysis of the requirement.
By the looks of it, this could have been written with Linux and mysql, and still had the same problems. These guys do not know WTF they are doing.
Cowboy programming, no matter what the tool, is always a bad idea.
And a corrupt Democrat in that situation would have done exactly the same thing. Neither side is better or worse than the other.
It is evidently up to us to keep those fools straight.
Demand code and process verification prior to the election. Run test after test, in public. Audit the test results. Lock the code and machines away, by a 3rd party, until actual use. Audit the results.
Still doesn't work. If you produce a paper copy, either for the voter, or to store away...then you have two different 'votes', cast by the same person. One on paper, and one in the db.
The paper says one thing, the database record may well say something else. And you, the voter, would never know. And what do you do in case of a close election? Have everyone bring back their hardcopies and count those?
Taking home a hardcopy then brings the situation of linking a voter to a vote. "If you expect to get paid this week, bring me your voting receipt that says you voted for Mr. X."
Or ones that will be overhead 'tomorrow'.
The 'excitement' is that the submitter just found this. And it's ver 5 instead of 4.
An aircraft launched missile has the advantage of being where you want it, when you want it.
Assuming the target is in low enough orbit, and fighter jet launching a missile can hit it in any orbit. A land based laser system has to wait until it is overhead.
The F-15 based ASAT system was not merely a proposal, but a developed and launch tested system.
Acquire the target, zoom up to about 80,000, launch. The missile takes over from there.
Within a few hours, you can hit a LEO sat in any orbital inclination around the globe.
Control
It's what we want, and it's what they want. (They being the entertainment industry, the media, whatever you want to call 'em)
However...the two parties want different types of control.
They want a return to before the 60's. When they had sole reign over what you saw, when you saw it, and how you paid for it. Want to watch a TV show? Have to wait until they air it, and watch the commercials they want you to. Movies? Have to pay for each and every viewing. On their schedule.
Lately, the scale has tipped more towards us. We can (sort of) watch what we want, when and how we want to.
A TV show? Sure...tape it, and watch it later. Or next year. But we've only 'paid' for it (by viewing the ads) once, if at all. And they've only goten paid once by the ad agency.
Movies? Get a pirate copy, and watch it over and over. Give it to your friends. Hell...give it to everyone.
They want that control back. They would like us to (again) pay every time we 'enjoy' some of their fine offerings. Pay either directly in money, or by watching ads.
Ultimately, I think they'd like for all the hard copy ability to go away. Pure streaming. Then, they can change things, insert new ads, get us to pay over and over and over again, for the same basic content.
If you watch a rerun on TV 2 years later, of course they don't have the same ads in it. They've gotten their consumers to pay twice for the same content. And they'd like that to continue.
With a VCR, PVR, whatever...a 'rerun' in your home will have the same ads in it as did the original. They've only gotten paid once.
The question is...who gets to decide who the control will reside with? CBS, et al, would obviously like to regain control over what, when, where, and how thay get paid. Currently, the control is (somewhat) in our hands. And getting more so everyday. Lets try to keep it that way.
Remember...we are not the consumers. We are the product. The ad agencies are the consumers.
Really? Look at a contrail some time after the aircraft goes by. It is quite diffuse.
There could have been localized winds, stirring up that part faster then the rest of it.
Leading to a wider appearing cloud of ice particles.
Which then reflect the setting sun.
Several people have said "Don't use your car!"
Others have said..."That's just not possible! I need to drive."
How about something in the middle?
Want to reduce rush hour traffic by an easy 10%? Find an alternate way to work twice a month. That's pretty much all it would take.
Not everyday. Not even every other day. If we could average alternate transport (carpool, bus, bike, whatever) twice a month, the problems would go down significantly.
I'm doing enough for about 5 of you (ride my bike 2-3 times a week).
What are you doing?
Excellent advice, however do not reuse empty milk jugs to keep the water in. The milk residue will eventually funk up the water.
Because the EULA is not for your benefit, but for theirs. If it were in plain English, the average user might actually be cognizant of the actual restrictions of what he is agreeing to, and then not do it.
Couched in legalese, you probably won't notice that the average EULA gives you, the consumer, no rights whatsoever, and protects the vendor from any action on your part.
how to protect themselves against legal action should any of the postings prove false.
And why, exactly, should they (whomever they may be) be immune from legal action should they post falsified data?
Check your sources before distributing some random bit of gossip.
You obviously do not have children.
Or, if you do, you do not trust them. And in return, they will not trust *you*.
Anybody else notice the large slowdown in downloading from emusic?
Evidently, everyone is d/ling as much as they can before the limits take effect.
I, for one, do not welcome our new Dimensional overlords.
$.25 is not unreasonable. It's the particulars of this plan that does kind of suck. You have a very low, hard limit of how many tracks you can d/l. And you're being charged that whether you use it or not.
There is no carryover of unused tracks.
Currently, there is no way to actually buy more. Either 40 tracks for $10, or 65 for $15. Unless you can and want to buy into the $50 plan.
emusic has a very eclectic catalog. Part of the draw there was being able to sample different artists at will. Now...you'd have to pick and choose very carefully what you wanted to spend your 40 tracks on.
It's 'bad' because now you get less for the same money.
I've been with them almost a year. My sub runs out in Nov. (Now...it's Nov 7 to be exact)
In that time, I've grabbed about 130 cd's. So maybe 12 cd's per month. 120 tracks on average. Often, I might go a month or two without anything, and then go get a bunch all at once.
With this new d/l limit, I'd have to cut back to 1/3. About 4 cd's per month, for the same price. And no month to month carryover of unused tracks.
Plus, now you'd have to be MUCH more careful about which tracks you actually d/l. Gone will be the concept of "just get the whole album". If I were to continue, I'd pick and choose each track so as to maximise my selections. Previous, if a few tracks on the album sucked...so what. It didn't cost anything extra.
but they are company trying to make a profit.
Right. My question is...were they making a profit before, or is it simply a case of the new owners wanting to make more profit? IMO, they are making a big mistake, and there will be a mass exodus fo current subscribers.
See ya emusic.
I wonder if their catalog will be changing/expanding.
If not...bye bye emusic. It was nice knowing ya.
A fixed gear is a different kettle of fish altogether. There is no freewheel. As the rear wheel turns, the pedals turn. All the time.
Folding bikes look to be about $70-$150 on eBay.
As with most things, there is a minimum level of acceptable quality. Buying a $70 folding bike from eBay is probably below that level.
Would you buy and drive a Yugo, just because it was cheap?