I helped write the original eVACS system. Forking the code for a US voting system is a nice idea, but probably won't be as helpful as you might like. Most of the complexity in the eVACS code is dealing with the ACT's Hare-Clark electoral system. That affects both the voting interface and the back end counting system. It even affects the system's whole architecture, because the votes have to all be recorded, then counted as a batch, rather than tallied as they are entered which is the obvious way to count a first-past-the-post US style election.
So looking at the system might yield some good ideas about how to organise the system (in particular how the sequence of voting and authentication is handled), but I don't think all that much code could be reused.
If normal people can't count it, it's too complex and too easily corrupted. If only.01 of 1% of the people can count the ballots and assure the quality, they can easily create a conspiracy(like Bush's luck that his brother was gov. of Florida.)
The original statement was misleading. With half an hour's work or so anyone can understand how to count votes using Hare-Clark. The paper counting process has always been observed by scrutineers from all parties who know how to apply the system. The idea of an open source electronic system is that party scrutineers can now check the code to ensure that it correctly applies the system.
Understanding why it gives sensible results is a bit harder (for some of the edge cases) but still well within the reach of most people with a bit of thought and some worked examples.
It's all open source, both the voting booth front end and the back end counting and data entry systems.
See http://evacs.samba.org/ for *prototype* code.
Voting is compulsory in Australia, but failing to vote won't land you in prison any more than a parking ticket would. There is a fairly nominal fine for failing to vote.
I've looked at this. It's not really suitalbe for this application though.
First, FREE is heavily geared towards internet voting - the only interesting parts of the implementation are involved in this. Here the voting is happening in ordinary polling places so the problems are quite different.
Secondly, FREE pretty much assumes first-past-the-post non-preferential voting. That makes it useless for the ACT both in terms of the front end used interface (it has no facility for listing preferences) and in terms of the tally system (it doesn't implement the Hare-Clark counting algorithm).
It is somewhat ironic that as far as I'm concerned, one of the most serious screwups in Aus. Govt. IT policy is the fact that several state governments have made outsourcing contracts which include provisions binding them to Microsoft for ludicrous lengths of time (20 years for the South Australian govt., I believe, via a contract with EDS). It boggles the mind that one could even contemplate signing an exclusive contract with a company for nearly as long or longer than that company has existed.
I've been looking quite a bit at the linux-wlan code which you can get from the Absolute Value systems page. Frankly, it's bloated beyond belief and in some places just plain broken (e.g. a jiffies based timeout with interrupts disabled). It has very much the look of code produced by someone being paid by the line. I'd rate its chances of getting into the standard kernel in anything like its present form as virtually nil.
It most certainly is point to point. There's no way you'd get the distances covered here with omnidirectional within the legal power limits. The Galaxy antennas aren't "satellite dishes" as such (Galaxy was a surface system), just directional microwave antennas.
Not really. In order to get the distances, you need to use fairly highly directional antennas. So the restaurant would have to be very well located, and then you'd have to get your dish in to line of sight of another antenna and carefully line it up.
Actually the Galaxy antennas in question do include a downconverter/preamplifier which needs to be ripped out before they can be used for transmission.
Galaxy wasn't a satellite system. It was a surface pay TV system based on microwave transmissions from Telstra tower on black mountain.
All that the air.net.au people are using are the now disused microwave antennas - they can often be collected free from people who used to have Galaxy. That makes a very cheap way of getting an ~18dBi antenna.
On a tangentially related note, some work has been done on stabilizing radioactive waste prior to burial or storage.
The CSIRO in Australia a few years ago developed a material called "synrock". As the name suggests, it's essentially a synthetic "rock" in which the waste is embedded. It's chemical and physical properties are supposed to be such that the waste won't move (well, not much). Presumably it would have properties somewhat similar to uranium ores.
Saying that we (Australia) have the population of New York spread out over the land area of the US is misleading. The figures are more-or-less correct, but to say we're "spread out" would be stretching things. Australia is one of the most urbanised countries in the world, with >75% of the population concentrated in the big(ish) cities (Sydney, Melbourne, Brisbane, Adelaide and Perth).
I don't think it's really accurate to describe the Australia card as being a similar scheme to this. It was a fair while time ago, so my recollection is rather fuzzy but I believe the Australia card was supposed to be a universal ID and index for government services. I think it was to be fairly similar in concept to the French ID card (which has been around for quite some time, and as far as I know is not massively unpopular).
While that certainly has privacy concerns it wasn't, unlike this scheme, about collecting data for the express purpose of selling market information. While having large government databases with personal information is a cause for concern, to my mind having similar privately held databases is much, much worse.
A "Linux kernel based" OS that's not just a GNU/Linux distribution is actually possible: You could take the kernel (unmodified) but replace all of userland with a different system. The new userland could be distributed under any license one pleased, including fully proprietary.
With a different C library, standard utilities and file system conventions you could have a thoroughly non-Unix-like OS, even with POSIX system calls under it all.
Of course, why you'd bother is another question. I'd be fairly doubtful myself as to whether this is what they've actually done.
I find it especially interesting that Katz invokes John Ralston Saul. Saul is very much in favour of (considered) regulation and certainly no libertarian (at least, not in the sense it's usually meant on the 'net).
It seems to me that better than banning all cookies attached to GIFs, browsers should have reasonable cookie handling options. So far I seen precisely one browser that appears to handle them vaguely sensibly, and that is Lynx (!).
When you are prompted to accept a cookie, as well as Yes and No options, you can indicate Always or Never to accept cookies from that domain. If netscape had this option, and kept your choices persistently (which Lynx doesn't for now) the matter is very simple. Just choose 'Never' the first time you see a doubleclick cookie. In addition Lynx allows you to view all currently held cookies and discard them at will or change Always/Never options by domain.
As far as I understand it, every single signatory to the Berne convention _except_ the US has first-to-file patents rather than first-to-invent. First-to-invent is a actually a rather bad idea. It sounds like its fairer to small inventors, but in practice it isn't.
The only way to prove you're first to invent (if someone else filed first) is to go to court. Hence first-to-invent ends up meaning whoever can last longest in a court battle. Even if you have good evidence that you did invent it first (and that's going to be difficult) you still have to get the cash together for a court challenge.
Under first-to-file, the small guy has at least a fighting chance to get to the patent office first.
Slashdot Mirror
I helped write the original eVACS system. Forking the code for a US voting system is a nice idea, but probably won't be as helpful as you might like. Most of the complexity in the eVACS code is dealing with the ACT's Hare-Clark electoral system. That affects both the voting interface and the back end counting system. It even affects the system's whole architecture, because the votes have to all be recorded, then counted as a batch, rather than tallied as they are entered which is the obvious way to count a first-past-the-post US style election.
So looking at the system might yield some good ideas about how to organise the system (in particular how the sequence of voting and authentication is handled), but I don't think all that much code could be reused.
The original statement was misleading. With half an hour's work or so anyone can understand how to count votes using Hare-Clark. The paper counting process has always been observed by scrutineers from all parties who know how to apply the system. The idea of an open source electronic system is that party scrutineers can now check the code to ensure that it correctly applies the system. Understanding why it gives sensible results is a bit harder (for some of the edge cases) but still well within the reach of most people with a bit of thought and some worked examples.
It's all open source, both the voting booth front end and the back end counting and data entry systems. See http://evacs.samba.org/ for *prototype* code.
Um, no.
Voting is compulsory in Australia, but failing to vote won't land you in prison any more than a parking ticket would. There is a fairly nominal fine for failing to vote.
I've looked at this. It's not really suitalbe for this application though.
First, FREE is heavily geared towards internet voting - the only interesting parts of the implementation are involved in this. Here the voting is happening in ordinary polling places so the problems are quite different.
Secondly, FREE pretty much assumes first-past-the-post non-preferential voting. That makes it useless for the ACT both in terms of the front end used interface (it has no facility for listing preferences) and in terms of the tally system (it doesn't implement the Hare-Clark counting algorithm).
Since voting in ACT elections is compulsory for residents anyway, this isn't terribly relevant.
Well, sort of compulsory, anyway - the fine for failing to vote is not large.
It is somewhat ironic that as far as I'm concerned, one of the most serious screwups in Aus. Govt. IT policy is the fact that several state governments have made outsourcing contracts which include provisions binding them to Microsoft for ludicrous lengths of time (20 years for the South Australian govt., I believe, via a contract with EDS). It boggles the mind that one could even contemplate signing an exclusive contract with a company for nearly as long or longer than that company has existed.
I've been looking quite a bit at the linux-wlan code which you can get from the Absolute Value systems page. Frankly, it's bloated beyond belief and in some places just plain broken (e.g. a jiffies based timeout with interrupts disabled). It has very much the look of code produced by someone being paid by the line. I'd rate its chances of getting into the standard kernel in anything like its present form as virtually nil.
It most certainly is point to point. There's no way you'd get the distances covered here with omnidirectional within the legal power limits. The Galaxy antennas aren't "satellite dishes" as such (Galaxy was a surface system), just directional microwave antennas.
Not really. In order to get the distances, you need to use fairly highly directional antennas. So the restaurant would have to be very well located, and then you'd have to get your dish in to line of sight of another antenna and carefully line it up.
Actually the Galaxy antennas in question do include a downconverter/preamplifier which needs to be ripped out before they can be used for transmission.
Galaxy wasn't a satellite system. It was a surface pay TV system based on microwave transmissions from Telstra tower on black mountain.
All that the air.net.au people are using are the now disused microwave antennas - they can often be collected free from people who used to have Galaxy. That makes a very cheap way of getting an ~18dBi antenna.
On a tangentially related note, some work has been done on stabilizing radioactive waste prior to burial or storage.
The CSIRO in Australia a few years ago developed a material called "synrock". As the name suggests, it's essentially a synthetic "rock" in which the waste is embedded. It's chemical and physical properties are supposed to be such that the waste won't move (well, not much). Presumably it would have properties somewhat similar to uranium ores.
Saying that we (Australia) have the population of New York spread out over the land area of the US is misleading. The figures are more-or-less correct, but to say we're "spread out" would be stretching things. Australia is one of the most urbanised countries in the world, with >75% of the population concentrated in the big(ish) cities (Sydney, Melbourne, Brisbane, Adelaide and Perth).
I don't think it's really accurate to describe the Australia card as being a similar scheme to this. It was a fair while time ago, so my recollection is rather fuzzy but I believe the Australia card was supposed to be a universal ID and index for government services. I think it was to be fairly similar in concept to the French ID card (which has been around for quite some time, and as far as I know is not massively unpopular).
While that certainly has privacy concerns it wasn't, unlike this scheme, about collecting data for the express purpose of selling market information. While having large government databases with personal information is a cause for concern, to my mind having similar privately held databases is much, much worse.
A "Linux kernel based" OS that's not just a GNU/Linux distribution is actually possible: You could take the kernel (unmodified) but replace all of userland with a different system. The new userland could be distributed under any license one pleased, including fully proprietary.
With a different C library, standard utilities and file system conventions you could have a thoroughly non-Unix-like OS, even with POSIX system calls under it all.
Of course, why you'd bother is another question. I'd be fairly doubtful myself as to whether this is what they've actually done.
Hear, Hear!
I find it especially interesting that Katz invokes John Ralston Saul. Saul is very much in favour of (considered) regulation and certainly no libertarian (at least, not in the sense it's usually meant on the 'net).
It seems to me that better than banning all cookies attached to GIFs, browsers should have reasonable cookie handling options. So far I seen precisely one browser that appears to handle them vaguely sensibly, and that is Lynx (!).
When you are prompted to accept a cookie, as well as Yes and No options, you can indicate Always or Never to accept cookies from that domain. If netscape had this option, and kept your choices persistently (which Lynx doesn't for now) the matter is very simple. Just choose 'Never' the first time you see a doubleclick cookie. In addition Lynx allows you to view all currently held cookies and discard them at will or change Always/Never options by domain.
As far as I understand it, every single signatory to the Berne convention _except_ the US has first-to-file patents rather than first-to-invent.
First-to-invent is a actually a rather bad idea. It sounds like its fairer to small inventors, but in practice it isn't.
The only way to prove you're first to invent (if someone else filed first) is to go to court. Hence first-to-invent ends up meaning whoever can last longest in a court battle. Even if you have good evidence that you did invent it first (and that's going to be difficult) you still have to get the cash together for a court challenge.
Under first-to-file, the small guy has at least a fighting chance to get to the patent office first.
IANAL, YMMV, standard disclaimers apply.