Identity
Filter
Visitors
(0) (0) (0) (0) (1) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (61) (0) (0) (0)Subscribers
(0) (0)
Highest posts
- 24/10 10:44 +2 Re: Software initiatives
- 23/10 17:05 +2 Re: Software initiatives
- 23/10 15:17 +2 Re: Software initiatives
- 23/10 14:59 +2 Re: Software initiatives
- 23/10 13:03 +2 Re: Software initiatives
- 24/10 12:18 +1 Re: Software initiatives
- 24/10 10:29 +1 Re: Software initiatives
- 24/10 10:24 +1 Re: Software initiatives
- 24/10 10:15 +1 Re: Software initiatives
- 24/10 00:58 +1 Re: Software initiatives
Latest posts
- 24/10 12:18 +1 Re: Software initiatives
- 24/10 10:44 +2 Re: Software initiatives
- 24/10 10:29 +1 Re: Software initiatives
- 24/10 10:24 +1 Re: Software initiatives
- 24/10 10:15 +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
Links
Blogs
Democracy
Search
[top-politics] Re: Software initiatives
-
On 10/20/06, echarp <emmanuel.charpentier@free.fr> wrote:
> Let me try to summarize: > * real life person is public, many secret actions > * real life person is secret, many public actions > > > And I don’t se why it would be easier. > > In the first case, there are many secrets for one person. In the second > there is only one secret per person. This is why I consider it simpler. > > Plus it allows to develop an online persona with a history and a > reputation.
I agree with this.
And i also agree that using only one software, even if the source of the software is public and tested by many, is a bad idea. Some bugs will not be found untill the have made their share of harm. The best way to protect against bugs is to simultanly use independent software.
+2
-
Unfortunately the simples programming isn’t good enough if secret votes shall be an option for individual voters.
I’ll start a new thread about that.Regarding different soft, sure, nothing stops the paralell use of different versions in order to find bugs and security issues. But one and only one will have to be the official when this system is in production.
There should be a very trusted board, with immediate recall, responsible as admins for this system.
+1
-
> Regarding different soft, sure, nothing stops the paralell use of > different versions in order to find bugs and security issues. > But one and only one will have to be the official when this system is > in production.
Planes and rockets use different softwares and technologies.
Why not also do the same with voting systems? Is that such a weird idea?
> There should be a very trusted board, with immediate recall, > responsible as admins for this system.
That board is a single point of failure. Brittle.
Better to let all interested parties in the system, like it’s already the case in most elections where you have many boards with many participants.
echarp – http://leparlement.org/security
+1
-
At 11:39 AM 10/21/2006, echarp wrote:
> > There should be a very trusted board, with immediate recall, > > responsible as admins for this system. > >That board is a single point of failure. Brittle.
If what you are building is a power structure, as distinct from a corruption-immune system like FA/DP — which maintains its relative immunity by leaving nearly all power at the member level — then you will almost certainly need a board, and officers. Yes, this is a possible failure point, no doubt about that, but the problems with boards happen when the members are not actively involved and there is no means for a member recognizing the problem to effectively and efficiently communicate with the rest of the members, the board can block it.
This is why I propose FA/DP as a communications structure, and non-FAs, perhaps with DP structure, in power structures. And that “power structure” might quite well elect and have the power of recall over a board. Which then hires and has the power of firing employees. Officers are employees.
It’s the corporate model, folks, been going for hundreds of years. Where it goes astray is where there are large numbers of uninvolved and uninformed shareholders, because the structure was designed for relatively small numbers of investors. And, you will note, it incorporates proxy voting. Because those investors would not have accepted less!
Delegable proxy is just an extension and not even a totally necessary one, as far as the actual power structure is concerned. If you have an FA/DP superstructure, the FA/DP proxy can recommend an appropriate legal proxy to the member. And that can be a standing appointment, revocable at any time. This is what a proxy is!
So it could be pretty safe, even with a board holding the immediate reins of power. As long as they cannot control communication and analysis among the members!
+1
-
> If what you are building is a power structure, as distinct from a > corruption-immune system like FA/DP — which maintains its relative > immunity by leaving nearly all power at the member level — then you > will almost certainly need a board, and officers. Yes, this is a > possible failure point, no doubt about that, but the problems with > boards happen when the members are not actively involved and there is > no means for a member recognizing the problem to effectively and > efficiently communicate with the rest of the members, the board can block it.
Basically, we don’t know what happens after democratic decisions are taken. Board, government, administrators, commission.
It is open to all possibilities.
But, during the vote, any single point of failure ought to be removed to strengthen the whole thing. Consensus is the tool of choice to agree on what the expressed decisions are (as is currently the case).
echarp – http://leparlement.org
+1
-
>Basically, we don’t know what happens after democratic decisions are taken. Board, government, administrators, commission.
It is open to all possibilities.
But, during the vote, any single point of failure ought to be removed to
strengthen the whole thing. Consensus is the tool of choice to agree onwhat the expressed decisions are (as is currently the case).
-Just like to add that what possibilities and rights a bord should have ofcourse have to very clearly specified and it shall only be more or less technical issues about the voting process in relation to the design intent, nothing more.
So his has to be written clearly in a regulation document, otherwise this board soon enough will contain old politicians seeing a way of preserving their own power instead of acting simple clerks, working more as servants to the voters.
+1
-
> > But, during the vote, any single point of failure ought to be > > removed to strengthen the whole thing. Consensus is the tool of > > choice to agree on what the expressed decisions are (as is currently > > the case). > > Just like to add that what possibilities and rights a bord should have > ofcourse have to very clearly specified and it shall only be more or > less technical issues about the voting process in relation to the > design intent, nothing more.
The voting system has to be defined for the voters to cast their vote. Anybody willing should be able to follow that through and, with the votes database, be able to come to the correct results.
The voting procedure should be defined… democratically! ;)
> So his has to be written clearly in a regulation document, otherwise > this board soon enough will contain old politicians seeing a way of > preserving their own power instead of acting simple clerks, working > more as servants to the voters.
Coming to a consensus is a matter of technicality. Otherwise it is open to all sorts of other matters.
That consensus should become a widespread habit, where many will check all results, and media would report their own. Then it’s a matter of reputation and outcomes.
echarp – http://leparlement.org/fr
+2
-
+1
+1
-
-
-
-
+1! Exactly my point!
With TOP applied on this board, no risk for corruption and misleading
of the users, the people.
+1
-
-
You are misunderstanding.
I’m not against different initiatives and developments. I’m only saying that for every given vote casting, there shall be one and only one soft counting and handling all votes casting. This is obvous need for having trust in the democratical process. 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 board I’m talking about is the only way to control that this is the case.
It would be the parallell to todays voting authorities.
+1
-
> 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
-
-
-
-
-
-
-
-
-