There were two topics going, and you confounded them.
On the topic of patenting mathematics, in any form, I wrote:
I don't think it is to society's advantage to make the bargin, and thus I oppose it.
To which you replied:
I think I understand what the problem is now. By "the concept is sound", I mean that it can be to a society's advantage, depending on various details such as how long the patent is effective. By objecting to it, what you mean appears to be that given this set of circumstances, you don't think it's helpful.
To reiterate, I find it hard to imagine a set of circumstances in which a society would be well advised to, even in part, give up the principle of freedom-of-thought to learn a few algorithms, any more than I think it would be a good idea to sell your vital organs to raise some quick cash.
On the topic of patents in general I wrote:
I think a small, poor society would be well advised to grant patents liberally, whereas I think a large, rich society should be much more parsimonious.
To which you replied:
Just to nitpick, then it doesn't appear to be a matter of principle, but relative weighing of costs and benefits.
The principled rejection was of patenting discoveries in natural law, mathematics, etc. The pragmatic reasoning was regarding patents of normal inventions, which I agree may be worthwhile but ought to be sharply curtailed from today's orgy of artificial-monopolys-as-indulgences.
Good. I think recognizing that we disagree is the key to this discussion.
The patentability of algorithms, as a concept, is sound.
I disagree, as previous stated.
It is material incentive to discover new algorithms, and in exchange for publication, the society grants a limited monopoly to exploit the discovery.
I agree that it is an incentive, but disagree that it is needed (by definition, a useful algorithm is worth discovering for its own sake--no additional incentive should be required). I would agree with your statement of the bargin society makes (though I might disagree that the monopoly is "limited" in most practical situations). But I don't think it is to society's advantage to make the bargin, and thus I oppose it.
The implementation, at least in the US, is of course deeply flawed.
Agreed.
I am not willing to give up even a small part of my right to do mathematics in order to induce them to do so
Well said, even though I disagree with the absolutist approach.
*smile* An absolutist appoach is hard to avoid on matters of principle--at least, if your principles mean anything to you. I suppose you could make the argument that some amount of thought police was good, but I'd have a hard time swollowing it.
Do you consider anything at all patentable, though?
Good question. I think a small, poor society would be well advised to grant patents liberally, whereas I think a large, rich society should be much more parsimonious. In the former, the value of inventions is high and the chance of losing them great, whereas in the later the incremental value is small and the chance that any one invention will be made many times is high.
I suppose I would like consider things patentable only if they were very valuable and very "non-obvious"; perhaps (if I were asked to redesign the patent system) I would recommend granting a small number of patents (say, a dozen or fewer a year) on devices of proven value in cases where it appeared that there was a real risk that the invention would be lost forever if not disclosed.
Too many incentives leads to madness; a state to which which we as a society are far closer than most of us care to admit.
-- MarkusQ
P.S. Or, to be contrarian, for the next hundred years or so I would be willing to grant patents to any device which was developed in geosyncronous orbit (or higher) covering use anywhere nearer the surface of the earth than the point of development.
I must be odd man out. I had one go bad on me (normal office environment). It got into a mode where it would slow way down, and packet sniffing (with etherial) showed that it was trying to "phone home" (a huge number of HTTP puts to an IP address that resolved to labs.linksys.com IIRC). Cycling power seemed to fix it for a day or two. Three others in the same environment seemed fine.
-- MarkusQ
P.S. I also have a Netgear (don't recall the model) that goes wonkers in lightning storms--though I can't say I blame it.
Take MP3 encoding, for example. It's obviously worth money to you...You don't seem to have read my post at all before responding...Just so it's perfectly clear to you, I'll reiterate.
Just so it's perfectly clear to you, I have read your posts but I disagree with you. Although it is obvious to you that MP3 encoding is worth money to me (though I have no idea why that should be obvious or on what basis you claim to know what I value and what I don't), it is not obvious to me that having someone implement one of the many possible audio compression schemes should be worth compromising what is to me a fundumental principle:
No one owns mathematics: no one owns algebra, no one owns logic, no one owns algorithms, no one owns group theory, no one owns liniar algebra,...
The patentability of algorithms is a recent notion that has been snuck into the patent laws by people who want to make a buck by playing on the greed and ignorance of the public. It sets a dangerous precedent and I object to it on principle. The primary consequence is not (as you imply) that people now disclose their mathematical discoveries when previously did not, but rather that whoever discovers a fact first gets to decide who can use the fact and who can not.
Yes, it would be handy for me if people who thought up clever algorithms told me about them. No, I am not willing to give up even a small part of my right to do mathematics in order to induce them to do so, nor conceed to them the roll of "gatekeeper" over even a small corner of Absolute Truth.
Historically, a patent is a legal device to let the government protect and encourage the inventor or discoverer from copycats. It is particularly necessary as we entered the industrial age, where any useful widget you can invent can probably be easily copied and mass produced by an immoral competitor.
The salient distinction between humans and other animals is the extent to which we learn from others, primarily by copying examples of successful behaviour. In order to assure that this copying can continue after the death of the inventor of a complex device many societies have entered into a bargin with the inventors: if you tell us everything we need to know in order to copy your invention we will refrain from doing so for a limited time. This is the origin of patents. Protecting the inventor is not the purpose, it's the bribe society offers to get what it wants. This deal is offered to inventors whose work is:
Original There's no point in society paying a premium for something that it already has.
Non-obvious Likewise, there's no point in buying something that you would have gotten anyway
Fully disclosed Remember, the goal is to enable eventual copying.
Demonstrated by a working example We aren't willing to pay for something that doesn't work.
These restrictions are put in place because the price we pay is so high--nothing less than collectively giving up a portion of that which makes us human.
I think few will argue that the discoverers (assuming you also don't like to use "invent" for algorithms) of certain algorithms deserve compensation.
Then you should get out more. Many people do argue just that, and claim that algorithms should not be protected by patents.
I think it is a very real danger that without patent protection, many useful discoveries will simply be buried deep within the biggest companies.
That's just nuts. Most of these "inventions" are so obvious that the main aplication of patents these days isn't to stop people who learned how to do it by reading the patent, but to chill inovation by bludeoning anyone who "invents" something similar.
I don't have a real problem with your ideals, but it's not reality yet, and therefore have little bearing on this discussion.
What do you mean, yet? What he was describing has been the state of affairs for the bulk of the history of patents. Only in the last few years has the nonsense of issuing patents on mathematics been given any credence whatsoever. For that matter, I'm not convinced that "the law as it stands" really supports the modern practice of issuing patents on algorithms.
Urgent nitice to whoever modded my post above "redundant":
Turn around at once! Since I posted over an hour before the post that made you think mine was redundant I can only conclude that you are experiencing time backwards! As I'm sure you know, this is very dangerous, especially if you don't realize that you're doing it before it's too late (or should that be "after it's too early"?). Anyway, turn back now before (after?) somebody gets (got?) hurt.
If not, don't. Liquid paper is not a good way to whiten your teeth. Normally I wouldn't think it neccessary to point that out, but since you also seemed to have missed the fact that:
Slashdot is not a good place to get medical advice
Don't just study them. Join them. Make a point of meeting people in the company, learning about the other products, customers, etc. Be friendly with everyone--the first person you get to know may turn out to be a great ally or they may turn out to be the universally lothed office jerk--so don't "choose sides" until you've been there a while. In short, act like a new hire rather than a member of an organ transplant just waiting for the T-cells to show up.
we (the quints) do receive a cut of the revenue we bring in from new installations of our software
Don't count on this unless there are specified minimums. I've seen a case that sounds exactly like what you are describing, where after a week or so of happy-happy, the new hires started getting assigned to help another team that was developing a replacement product, were told to add a data export feature to their product, etc. They wound up doing 90% of the work to migrate their old customers to the new owner's product (including helping the sales team when it came to that), and then they were all (almost all? --this was several years ago) let go. From the day of purchace, there was never a single new installation of their product.
Of course, I have seen it work out well (in one case the purchaser was a hardware vendor and the company they bought wound up being a very technical vertical market sales team, getting paid their old base pay, plus (in some cases) commissions, plus the bonus based on their software); it can work out well; but I would not count on it.
-- MarkusQ
P.S. Be especially suspicious if the cut you get when your software sells seems too good to be true.
What makes you think anyone would care about oil on Mars...and more than they care about the coal, frozen chunks of natural gas, iron (not iron ore, but actual iron), gold, titanium, etc. that are floating around where they'd be even easier to get at than oil on Mars?
The only thing I can figure: most planners / politicians don't really believe in space, any more than they believe in Santa's workshop, history, or the constitution of the country that they "serve". These sorts things are all handy to have in your bag of speach writting tricks but that's about it.
This problem has already been solved. You should be searching under more general terms (e.g. "fair scheduling algorithms") rather than problem specific terms like "pager rotation," that's all.
For example, let's say you have N people working (and all are interchaingable, to start with). That means that each of them should be on call for K = 1000/N milliseconds out of every second (on average). Provided there are less than 500 people to be scheduled, you can accomplish this by rounding K to an integer (for the case where there are more that 500 people to be scheduled, either schedule them for one millisecond each, or go to a finer grained time-base). One important point to remember is that you must resource lock the call to the person in << K ms to avoid race conditions (which can garble text messages and result in an annoying high-pitched noise if two or more people try to return the call simultainiously and get multiplexed--
One, almost axiomatic right now is that even if we colonize space, we could never afford to lift any significant fraction of humanity off the surface
With solar power (at the top), beanstalks, and clever scheduling policy (in a nutshell: young women get priority) we could get everybody off within a generation or two. With automated canister city factories & asteriods to play with we could even have a nice place to put them.
The problem isn't energy. The problem is that you could never implement such a scheme, even as an evacuation, given how mule-headed people are.
Sci-Fi and sci-fi games often incorporate...ships with ports of call on a myriad planets...I was toying with the physics of propelling such ships and their cargos out of a gravity well and realized the astronomical amounts of power it would take to do it (not to mention interstellar travel).
Not so fast. You don't need that much energy to get from the bottom of one gravity well to the bottom of another, provided you can swap momentum around. There are a number of schemes along the lines of the cable cars that harvest energy from cars going down (and momentum from stopping cars) and feed it to cars going up / accellerating.
There is no such thing as a free lunch, but often the problem isn't the cost of the lunch but of all the stuff you wastefully throw away while eating it.
CDs will last at least as long as the average paper archive
Paper can easily last a hundred years (I have a number of books from the late 1800's & early 1900's); IIRC the typical MTF for CDs is on the order of 20 years, and can be as low as 5.
Don't you just love it when someone uses the term short and not really know what it means?
He shorted the jerks at $11. I say good for him, I shorted them at $9.60 and at $10 & change. What this means is that we borrowed shares of SCOX through our brokers and sold them at the stated price, in the expectation that by the time we have to return the borrowed shares the price will have fallen and we'll be able to replace them for much less. This is called shorting. I'm happy with my position (they can't keep up this ballony forever); I'd be even happier with his. So, what's your objection to the term "short" in this context?
numerous archives from seminal cases from the 50's, 60's and 70's. They are available in MP3 format
Wow! That boggles my mind! I want to know who had the foresight to record the cases in MP3 format way back in the 1950s. And even more importantly, I'd like to know if they have any investment tips to share...
What's more, the type of fix you describe is quite complicated - you'd have to be able to make the system perform sorts and multiple counts, and somehow cover this up.
No, it would be a piece of cake. Have the system pull the ballots to be counted "at random" from the set of uncounted ballots, using the usual "flop the last ballot into the open slot and decrement the counter" trick. Then let the "random" number generator see the ballots and favour ballots that rank your pet candidate / party low by generating multiple random numbers (say, ten) and choosing the one that ranks your candidate the lowest. Since the routine is returning an arbitrarily selected random number it should have the same (or even better) statistics than the un-fudged random number generator. By giving the variables involved names like "entropy_pool_index" and "bits_of_entropy_remaining" and throwing in comments about the published (paper only) source for the algorithm, you might even be able to get it by a code review. But you won't need to go to all this trouble because...
No one outside your organization gets to see the source code!
This would be a lot easier than stuffing the ballot box, and you could even print out an auditable record of how the tally was reached.
-- MarkusQ
Firstly each voter can vote in order of preference for every candidate - For example say there are 10 candidates for three seats (my case last election) You can vote in order of 1 to 10.
PR works by counting first how many ballots are cast, dividing by some ammount (IIRC Number of seats + 1). This is set as the "quota". Then counting takes place. Once a candiate reaches the quota they are deemed elected. Then the amount of votes over the quota is distributed to the other candates, going on the next choice of the voters concerned.
This is a complicated system and electronic counting would be an advantage
It sure would be. If you don't like the winner on the first count, just try again but count the ballots in a different order. Or even just sort them to give your guy the maximum chance of getting in (e.g. count ballots where he was the last choice first).
Both of these ways of fixing such an ellection would be impractical in a paper system, but quite easy in an electronic black box.
Slashdot Mirror
There were two topics going, and you confounded them.
On the topic of patenting mathematics, in any form, I wrote:
To which you replied: To reiterate, I find it hard to imagine a set of circumstances in which a society would be well advised to, even in part, give up the principle of freedom-of-thought to learn a few algorithms, any more than I think it would be a good idea to sell your vital organs to raise some quick cash.On the topic of patents in general I wrote:
To which you replied: The principled rejection was of patenting discoveries in natural law, mathematics, etc. The pragmatic reasoning was regarding patents of normal inventions, which I agree may be worthwhile but ought to be sharply curtailed from today's orgy of artificial-monopolys-as-indulgences.-- MarkusQ
I disagree.
Good. I think recognizing that we disagree is the key to this discussion.
The patentability of algorithms, as a concept, is sound.
I disagree, as previous stated.
It is material incentive to discover new algorithms, and in exchange for publication, the society grants a limited monopoly to exploit the discovery.
I agree that it is an incentive, but disagree that it is needed (by definition, a useful algorithm is worth discovering for its own sake--no additional incentive should be required). I would agree with your statement of the bargin society makes (though I might disagree that the monopoly is "limited" in most practical situations). But I don't think it is to society's advantage to make the bargin, and thus I oppose it.
The implementation, at least in the US, is of course deeply flawed.
Agreed.
Well said, even though I disagree with the absolutist approach.
*smile* An absolutist appoach is hard to avoid on matters of principle--at least, if your principles mean anything to you. I suppose you could make the argument that some amount of thought police was good, but I'd have a hard time swollowing it.
Do you consider anything at all patentable, though?
Good question. I think a small, poor society would be well advised to grant patents liberally, whereas I think a large, rich society should be much more parsimonious. In the former, the value of inventions is high and the chance of losing them great, whereas in the later the incremental value is small and the chance that any one invention will be made many times is high.
I suppose I would like consider things patentable only if they were very valuable and very "non-obvious"; perhaps (if I were asked to redesign the patent system) I would recommend granting a small number of patents (say, a dozen or fewer a year) on devices of proven value in cases where it appeared that there was a real risk that the invention would be lost forever if not disclosed.
Too many incentives leads to madness; a state to which which we as a society are far closer than most of us care to admit.
-- MarkusQ
P.S. Or, to be contrarian, for the next hundred years or so I would be willing to grant patents to any device which was developed in geosyncronous orbit (or higher) covering use anywhere nearer the surface of the earth than the point of development.
I must be odd man out. I had one go bad on me (normal office environment). It got into a mode where it would slow way down, and packet sniffing (with etherial) showed that it was trying to "phone home" (a huge number of HTTP puts to an IP address that resolved to labs.linksys.com IIRC). Cycling power seemed to fix it for a day or two. Three others in the same environment seemed fine.
-- MarkusQ
P.S. I also have a Netgear (don't recall the model) that goes wonkers in lightning storms--though I can't say I blame it.
Take MP3 encoding, for example. It's obviously worth money to you...You don't seem to have read my post at all before responding...Just so it's perfectly clear to you, I'll reiterate.
Just so it's perfectly clear to you, I have read your posts but I disagree with you . Although it is obvious to you that MP3 encoding is worth money to me (though I have no idea why that should be obvious or on what basis you claim to know what I value and what I don't), it is not obvious to me that having someone implement one of the many possible audio compression schemes should be worth compromising what is to me a fundumental principle:
The patentability of algorithms is a recent notion that has been snuck into the patent laws by people who want to make a buck by playing on the greed and ignorance of the public. It sets a dangerous precedent and I object to it on principle. The primary consequence is not (as you imply) that people now disclose their mathematical discoveries when previously did not, but rather that whoever discovers a fact first gets to decide who can use the fact and who can not.Yes, it would be handy for me if people who thought up clever algorithms told me about them. No, I am not willing to give up even a small part of my right to do mathematics in order to induce them to do so, nor conceed to them the roll of "gatekeeper" over even a small corner of Absolute Truth.
-- MarkusQ
Historically, a patent is a legal device to let the government protect and encourage the inventor or discoverer from copycats. It is particularly necessary as we entered the industrial age, where any useful widget you can invent can probably be easily copied and mass produced by an immoral competitor.
The salient distinction between humans and other animals is the extent to which we learn from others, primarily by copying examples of successful behaviour. In order to assure that this copying can continue after the death of the inventor of a complex device many societies have entered into a bargin with the inventors: if you tell us everything we need to know in order to copy your invention we will refrain from doing so for a limited time. This is the origin of patents. Protecting the inventor is not the purpose, it's the bribe society offers to get what it wants. This deal is offered to inventors whose work is:
- Original There's no point in society paying a premium for something that it already has.
- Non-obvious Likewise, there's no point in buying something that you would have gotten anyway
- Fully disclosed Remember, the goal is to enable eventual copying.
- Demonstrated by a working example We aren't willing to pay for something that doesn't work.
These restrictions are put in place because the price we pay is so high--nothing less than collectively giving up a portion of that which makes us human.I think few will argue that the discoverers (assuming you also don't like to use "invent" for algorithms) of certain algorithms deserve compensation.
Then you should get out more. Many people do argue just that, and claim that algorithms should not be protected by patents.
I think it is a very real danger that without patent protection, many useful discoveries will simply be buried deep within the biggest companies.
That's just nuts. Most of these "inventions" are so obvious that the main aplication of patents these days isn't to stop people who learned how to do it by reading the patent, but to chill inovation by bludeoning anyone who "invents" something similar.
-- MarkusQ
I don't have a real problem with your ideals, but it's not reality yet, and therefore have little bearing on this discussion.
What do you mean, yet? What he was describing has been the state of affairs for the bulk of the history of patents. Only in the last few years has the nonsense of issuing patents on mathematics been given any credence whatsoever. For that matter, I'm not convinced that "the law as it stands" really supports the modern practice of issuing patents on algorithms.
-- MarkusQ
Urgent nitice to whoever modded my post above "redundant":
Turn around at once! Since I posted over an hour before the post that made you think mine was redundant I can only conclude that you are experiencing time backwards! As I'm sure you know, this is very dangerous, especially if you don't realize that you're doing it before it's too late (or should that be "after it's too early"?). Anyway, turn back now before (after?) somebody gets (got?) hurt.
Only thinking of the public's safety...
-- MarkusQ
I was wondering about the security implications of this until I read that:
So now it's not the security implications I'm wondering about.
-- MarkusQ
Have you tried liquid paper?
If not, don't. Liquid paper is not a good way to whiten your teeth. Normally I wouldn't think it neccessary to point that out, but since you also seemed to have missed the fact that:
I thought that it might be worth mentioning.-- MarkusQ
Don't just study them. Join them. Make a point of meeting people in the company, learning about the other products, customers, etc. Be friendly with everyone--the first person you get to know may turn out to be a great ally or they may turn out to be the universally lothed office jerk--so don't "choose sides" until you've been there a while. In short, act like a new hire rather than a member of an organ transplant just waiting for the T-cells to show up.
-- MarkusQ
we (the quints) do receive a cut of the revenue we bring in from new installations of our software
Don't count on this unless there are specified minimums. I've seen a case that sounds exactly like what you are describing, where after a week or so of happy-happy, the new hires started getting assigned to help another team that was developing a replacement product, were told to add a data export feature to their product, etc. They wound up doing 90% of the work to migrate their old customers to the new owner's product (including helping the sales team when it came to that), and then they were all (almost all? --this was several years ago) let go. From the day of purchace, there was never a single new installation of their product.
Of course, I have seen it work out well (in one case the purchaser was a hardware vendor and the company they bought wound up being a very technical vertical market sales team, getting paid their old base pay, plus (in some cases) commissions, plus the bonus based on their software); it can work out well; but I would not count on it.
-- MarkusQ
P.S. Be especially suspicious if the cut you get when your software sells seems too good to be true.
What makes you think anyone would care about oil on Mars...and more than they care about the coal, frozen chunks of natural gas, iron (not iron ore, but actual iron), gold, titanium, etc. that are floating around where they'd be even easier to get at than oil on Mars?
The only thing I can figure: most planners / politicians don't really believe in space, any more than they believe in Santa's workshop, history, or the constitution of the country that they "serve". These sorts things are all handy to have in your bag of speach writting tricks but that's about it.
-- MarkusQ
For example, let's say you have N people working (and all are interchaingable, to start with). That means that each of them should be on call for K = 1000/N milliseconds out of every second (on average). Provided there are less than 500 people to be scheduled, you can accomplish this by rounding K to an integer (for the case where there are more that 500 people to be scheduled, either schedule them for one millisecond each, or go to a finer grained time-base). One important point to remember is that you must resource lock the call to the person in << K ms to avoid race conditions (which can garble text messages and result in an annoying high-pitched noise if two or more people try to return the call simultainiously and get multiplexed--
Hot damn, my run just finished.
G'night all...
-- MarkusQ
One, almost axiomatic right now is that even if we colonize space, we could never afford to lift any significant fraction of humanity off the surface
With solar power (at the top), beanstalks, and clever scheduling policy (in a nutshell: young women get priority) we could get everybody off within a generation or two. With automated canister city factories & asteriods to play with we could even have a nice place to put them.
The problem isn't energy. The problem is that you could never implement such a scheme, even as an evacuation, given how mule-headed people are.
-- MarkusQ
Not so fast. You don't need that much energy to get from the bottom of one gravity well to the bottom of another, provided you can swap momentum around. There are a number of schemes along the lines of the cable cars that harvest energy from cars going down (and momentum from stopping cars) and feed it to cars going up / accellerating.
There is no such thing as a free lunch, but often the problem isn't the cost of the lunch but of all the stuff you wastefully throw away while eating it.
-- MarkusQ
CDs will last at least as long as the average paper archive
Paper can easily last a hundred years (I have a number of books from the late 1800's & early 1900's); IIRC the typical MTF for CDs is on the order of 20 years, and can be as low as 5.
-- MarkusQ
Don't you just love it when someone uses the term short and not really know what it means?
He shorted the jerks at $11. I say good for him, I shorted them at $9.60 and at $10 & change. What this means is that we borrowed shares of SCOX through our brokers and sold them at the stated price, in the expectation that by the time we have to return the borrowed shares the price will have fallen and we'll be able to replace them for much less. This is called shorting. I'm happy with my position (they can't keep up this ballony forever); I'd be even happier with his. So, what's your objection to the term "short" in this context?
-- MarkusQ
an ESRB press release
Wow! So human cloning is finally a reality?!
-- MarkusQ (aka MQR)
P.S. *sigh* They aren't giving out The Most Obscure Joke Of The Year Award anymore, are they?
The thing is, a fuel cell will do the same thing with _zero_ moving parts...
Or on the order of 10^23 moving parts, depending on how you look at it.
-- MarkusQ
With this product, there's no argument, because you can't argue with a box
Yeah, right.
Reminds me of that old Country-Western standard "Mamas, don't let your babies grow up to be black hats."
-- MarkusQ
With luck, someone who does have points will jump in...
-- MarkusQ
numerous archives from seminal cases from the 50's, 60's and 70's. They are available in MP3 format
Wow! That boggles my mind! I want to know who had the foresight to record the cases in MP3 format way back in the 1950s. And even more importantly, I'd like to know if they have any investment tips to share...
-- MarkusQ
What's more, the type of fix you describe is quite complicated - you'd have to be able to make the system perform sorts and multiple counts, and somehow cover this up.
No, it would be a piece of cake. Have the system pull the ballots to be counted "at random" from the set of uncounted ballots, using the usual "flop the last ballot into the open slot and decrement the counter" trick. Then let the "random" number generator see the ballots and favour ballots that rank your pet candidate / party low by generating multiple random numbers (say, ten) and choosing the one that ranks your candidate the lowest. Since the routine is returning an arbitrarily selected random number it should have the same (or even better) statistics than the un-fudged random number generator. By giving the variables involved names like "entropy_pool_index" and "bits_of_entropy_remaining" and throwing in comments about the published (paper only) source for the algorithm, you might even be able to get it by a code review. But you won't need to go to all this trouble because...
No one outside your organization gets to see the source code!
This would be a lot easier than stuffing the ballot box, and you could even print out an auditable record of how the tally was reached. -- MarkusQ
Firstly each voter can vote in order of preference for every candidate - For example say there are 10 candidates for three seats (my case last election) You can vote in order of 1 to 10.
PR works by counting first how many ballots are cast, dividing by some ammount (IIRC Number of seats + 1). This is set as the "quota". Then counting takes place. Once a candiate reaches the quota they are deemed elected. Then the amount of votes over the quota is distributed to the other candates, going on the next choice of the voters concerned.
This is a complicated system and electronic counting would be an advantage
It sure would be. If you don't like the winner on the first count, just try again but count the ballots in a different order. Or even just sort them to give your guy the maximum chance of getting in (e.g. count ballots where he was the last choice first).
Both of these ways of fixing such an ellection would be impractical in a paper system, but quite easy in an electronic black box.
-- MarkusQ
There is only one physical security system worth squat (IMHO): a single door and some old, cynical guy with a gun.
-- MarkusQ