Identity
Filter
Visitors
(0) (1) (0) (0) (1) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (61) (0) (0) (0)Subscribers
(0) (0)
Highest posts
- 23/10 17:05 +2 Re: Software initiatives
- 23/10 15:17 +2 Re: Software initiatives
- 23/10 14:59 +2 Re: Software initiatives
- 24/10 10:29 +1 Re: Software initiatives
- 24/10 00:58 +1 Re: Software initiatives
- 24/10 00:57 +1 Re: Software initiatives
- 23/10 23:27 +1 Re: Software initiatives
- 23/10 17:40 +1 Re: Software initiatives
- 23/10 17:38 +1 Re: Software initiatives
- 23/10 17:32 +1 Re: Software initiatives
Latest posts
- 24/10 10:29 +1 Re: Software initiatives
- 24/10 00:58 +1 Re: Software initiatives
- 24/10 00:57 +1 Re: Software initiatives
- 23/10 23:27 +1 Re: Software initiatives
- 23/10 17:40 +1 Re: Software initiatives
- 23/10 17:38 +1 Re: Software initiatives
- 23/10 17:32 +1 Re: Software initiatives
- 23/10 17:27 +1 Re: Software initiatives
- 23/10 17:05 +2 Re: Software initiatives
- 23/10 16:48 +1 Re: Software initiatives
Links
Blogs
Democracy
Search
[top-politics] Re: Software initiatives
-
> You are misunderstanding.
Maybe. But I think I do understand your point. Just I disagree with it.
> I’m not against different initiatives and developments.
Cool.
> I’m only saying that for every given vote casting, there shall be one > and only one soft counting and handling all votes casting.
And I’m comparing it with planes and rockets, where there are multiple softwares.
Do you see my point?
> This is obvous need for having trust in the democratical process.
Trust should be obtained through verifiability, transparency.
You have a database of votes, which can be replicated quite easily. You let everybody obtain it, you let everybody count.
Wouldn’t that be much better to obtain trust? Isn’t verifying everything by yourself much better than having a central organism whom you’ll have to trust?
> Just as we today have only one way of haveing our votes in elections > counted, by a predetermined group of people handling the ballot boxes > etc etc.
The group is not predetermined (in France anyway). There is generally one way to count, but I’m not sure there could not be others. I could try proposing it next time I do the counting.
The trouble with physical votes, is that you can’t copy it and process it multiple times in parallel.
But using internet you can.
> The board I’m talking about is the only way to control that this is the > case.
This screams of “single point of failures”. In engineering, you want to remove them.
echarp – http://leparlement.org
+2
-
>And I’m comparing it with planes and rockets, where there are multiple softwares.
Do you see my point?
-Well, no. In a 747 there is one and only one soft controlling the right rudder of the wing as response to the signals from cockpit. If there where several and one where coming to a different conclusion, a light might go on in cockpit.
But still there should be only one soft doing the job until the pilot says otherwise. (Written by a non-avation engineer..)>You have a database of votes, which can be replicated quite easily. You let everybody obtain it, you let everybody count.
Wouldn’t that be much better to obtain trust? Isn’t verifying everything
by yourself much better than having a central organism whom you’ll haveto trust?
-Sure, but still only one as regarded the choosen one. The others might show errors in number one which will improve for next time or restart a new vote.
The alternative you are suggesting, that we rely on mr Johns new fancy soft, coming to a new “obvious” result is not to build trust for internet democracy.>The group is not predetermined (in France anyway). There is generally one way to count, but I’m not sure there could not be others. I could try proposing it next time I do the counting.
-The group can be choosen out of people of the street or whatever, but their voting algortihm + method of counting will have to be predetermined. =Parallell servers running ONE soft, not many If there where money enogh, there could be other groups conting differently and applying other algos but that would have been for comparison and verification, not for legal purpose.
>> The board I’m talking about is the only way to control that this is the case.
>This screams of “single point of failures”. In engineering, you want to remove them.
-Well, what you are suggesting seems like wild west in comparison.. Who will judge which soft that counted correctly this very time? A new vote about that or??
There must be someone in charge for the pratical of running votes/elections just as for the rudder controls in a plane, this is basic and not an invention of mine.
+1
-
> > And I’m comparing it with planes and rockets, where there are > > multiple softwares. > > > > Do you see my point? > > Well, no. In a 747 there is one and only one soft controlling the > right rudder of the wing as response to the signals from cockpit.
Redundancy.
Modern aircrafts which rely on a software controller ensure that not one software fault will bring down the whole thing.
http://en.wikipedia.org/wiki/Aircraft_flight_control_systems :
> Therefore virtually all fly-by-wire systems are triply or quadruply > redundant: they have three or four computers in parallel, and three or > four separate wires to each control surface. If one or two computers > crash, the others continue working. In addition most early digital > fly-by-wire aircraft also had an analog electric, mechanical or > hydraulic backup control system.
Do you now see my point?
> The alternative you are suggesting, that we rely on mr Johns new fancy > soft, coming to a new “obvious” result is not to build trust for > internet democracy.
Who is to rely on mr Johns? The point is not to rely on one element.
> > > The board I’m talking about is the only way to control that this > > > is the case. > > > > This screams of “single point of failures”. In engineering, you > > want to remove them. > > Well, what you are suggesting seems like wild west in comparison..
It is not the wild west, but good old engineering practices.
Redundancy. Verifiability.
> Who will judge which soft that counted correctly this very time?
You don’t judge, you go see the person reporting a different result and try to come to an agreement.
There are different steps for that.
Is the votes database the same?
How are votes deciphered?
What request is used to calculate the result?
If our current voting procedure allow every party to participate and come to a consensus, why not do that again but using software and internet?
> There must be someone in charge for the pratical of running > votes/elections just as for the rudder controls in a plane, this is > basic and not an invention of mine.
The demo is in charge! ;-)
echarp – http://leparlement.org/security
+1
-
>Modern aircrafts which rely on a software controller ensure that not one software fault will bring down the whole thing.
http://en.wikipedia.org/wiki/Aircraft_flight_control_systems :
> Therefore virtually all fly-by-wire systems are triply or quadruply > redundant: they have three or four computers in parallel, and three or > four separate wires to each control surface. If one or two computers > crash, the others continue working. In addition most early digital > fly-by-wire aircraft also had an analog electric, mechanical or > hydraulic backup control system.
Do you now see my point?
-Well, OK, forgot that there is so much redundance in a plane.., but, Boeing would newer sell a plane without testing and certifying all these redundant programs.This is my point, Boeing acts here as a trusted board sorting out different redundant.
They don’t leave software development to opensource community and just bring them in as another redundance improver.>You don’t judge, you go see the person reporting a different result and try to come to an agreement.
-I’m fine with that and such parallell running can be used. But. for legal resons we cannot see this as test environment once using it for making laws etc.
In that stage there must be a core soft with as little bugs as possible.
All problems with it and better solutions found by competing soft shall ofcourse be reported and afterwards corrected/improved. But until a fix/improvemnet is fully tested and found OK, the old soft must be the case.
And there must be somebody in charge for this process, just as in Boeing.Otherwise no one will sit down at takeoff anymore..
+1
-
On 10/23/06, MG <magnusgus@gmail.com> wrote:
> -I’m fine with that and such parallell running can be used. > But. for legal resons we cannot see this as test environment once using > it for making laws etc. > In that stage there must be a core soft with as little bugs as > possible.
You seem to assume the bugs of a groups of parallel softwares is all bugs existing in any of the softwares. But infact the bugs of the group is only the ones existing in all softwares. Do you see why it makes total sence to use not one but a group of parallel software for a prodution environment (making laws)?
+2
-
But how shall a two result situation be handled when we have an
important vote?
And who will handle it?
+1
-
On 10/23/06, MG <magnusgus@gmail.com> wrote:
> > > But how shall a two result situation be handled when we have an > important vote? > And who will handle it?
You could use a selected board for that if you want to :)
+1
-
On Mon, Oct 23, 2006 at 05:36:54PM +0200, ketty wrote:
> On 10/23/06, MG <magnusgus@gmail.com> wrote: > > But how shall a two result situation be handled when we have an > > important vote? > > And who will handle it? > > You could use a selected board for that if you want to :)
Or try to come to a consensus.
As already happens in all voting offices.
echarp – http://leparlement.org/security
+1
-
>> You could use a selected board for that if you want to :)
-See?
>Or try to come to a consensus.
As already happens in all voting offices.
-Is it? And what even if so, if we call a discussion of a badly written or torn ballot a “software problem”, isn’t the people in that voting office a trusted board?
+1
-
> > > You could use a selected board for that if you want to :) > > See?
Of course you could, but it is bad engineering.
> > Or try to come to a consensus. > > > > As already happens in all voting offices. > > Is it? And what even if so, if we call a discussion of a badly written > or torn ballot a “software problem”, isn’t the people in that voting > office a trusted board?
They are not a trusted board.
They are all willing people and parties. They reach a consensus.
echarp – http://leparlement.org
+1
-
-
-
-
-
> > They don’t leave software development to opensource community and > > just bring them in as another redundance improver.
There are plenty of ways to make software. But, for verifiability there is no need for a complex and complete thing, as serge stated, a spreadsheet should do the trick.
> > I’m fine with that and such parallell running can be used. > > But. for legal resons we cannot see this as test environment once using > > it for making laws etc.
We already use a decentralized and consensual system (polling stations). Where would the laws conflict with how the laws are made?
The entity responsible for votes, in a democracy, is the demos… Isn’t that quite logical? :-)
> > In that stage there must be a core soft with as little bugs as > > possible. > > You seem to assume the bugs of a groups of parallel softwares is all bugs > existing in any of the softwares. But infact the bugs of the group is only the > ones existing in all softwares. Do you see why it makes total sence to use > not one but a group of parallel software for a prodution environment (making > laws)?
+1
echarp – http://leparlement.org/security
+1
-
-
-
-
Hey,
I think your disagreement (echard and Magnus) stems from the voting methods you advocate so that in essence I believe your arguments actually complement each other. In the case of AD’s continous voting approach, there is a need for a much more specific piece of soft to be integrated in the system as it determines the closing of a vote dynamically. This need for integration and crosstalk is absent from echarp’s approach as he is thinking of vote verification after the fact (if I am understanding both your arguments correctly), which indeed could be done with something as simple as an excel spreadsheet populated with the voting results.
One could imagine that different counting pieces of code could be certified by delegates handling voting methods in the system (trustworthiness solved, Magnus?) and used in parallel so as to ensure that should one counting system fail or be compromised, other systems may show the discrepancy and raise a red flag (satisfying mode of operation, echarp?). This way you would have the certification needed for results to be trustworthy while enjoying the added security of having several counting options, discrepancies in which would be a reasonnable basis for investigation into a possible manipulation or irregularity.
Best regards,
Serge
+2
-
On Mon, Oct 23, 2006 at 03:19:12PM +0200, Serge wrote:
> I think your disagreement (echard and Magnus) stems from the voting > methods you advocate so that in essence I believe your arguments > actually complement each other.
Hopefully.
> In the case of AD’s continous voting approach, there is a need for a > much more specific piece of soft to be integrated in the system as it > determines the closing of a vote dynamically.
I also advocate continuous voting. But then I don’t consider that votes need to be closed.
> This need for integration and crosstalk is absent from echarp’s > approach as he is thinking of vote verification after the fact (if I > am understanding both your arguments correctly), which indeed could be > done with something as simple as an excel spreadsheet populated with > the voting results.
+1
“After” is verifiability. But it can and should occur at all time, in real time.
With luck every voter could be able to look up the results on different nodes, and see if his personal votes are correctly accounted for.
> One could imagine that different counting pieces of code could be > certified by delegates handling voting methods in the system > (trustworthiness solved, Magnus?) and used in parallel so as to ensure > that should one counting system fail or be compromised, other systems > may show the discrepancy and raise a red flag (satisfying mode of > operation, echarp?).
The mode of operation… I don’t have anything precise in mind. The point is just to have as many verification nodes as possible, then a consensus to define the result.
If that was to work properly, that system would organise in real time and wouldn’t need anything formal.
> This way you would have the certification needed for results to be > trustworthy while enjoying the added security of having several > counting options, discrepancies in which would be a reasonnable basis > for investigation into a possible manipulation or irregularity.
Investigation of course, and deliberation, conviction. Using computers and the simplest procedures possible, it should be possible to quickly determine what went wrong and how to correct it.
> Best regards, > > Serge
Sincèrement
echarp – http://leparlement.org/security
+1
-
-
-
+1
Public voting records allow replicability of voting results in a simple format which in turn allows everyone to verify the vote count for themselves (and that their vote is what they did cast and hasn’t been altered)… This is simpe, straightforward and embodies the ideal of Transparent, Open and Public. Echarp I agree wholeheartedly with your approach.
Best regards,
Serge
+2
-
+1
Ofcourse!
+1
-
-
