Re: Internet voting is about more than just software -- SOLVED?
Introduction
------------
jose lorenzo wrote:
> A democratic voting system whose results are intended
> to be binding is a great concept, but whose problems
> are solved by this?
>
> It is assumed that people on the web are free to roam
> and move from one region to another at will. In real
> life, if I do not like a law that is passed in the
> region I live, I have little choice but to live with
> it.
I think one of the GNU Herds goals is supply to the community a sane site
where there is not need to move out. Join communities no split. Members
having to move out due to organisation issues is against it. I think the
main intention of the voting system is help members control the management
team. It can be used too to get opinions about ideas, etc.
* binding vote
* opinion vote
> A voting system can provide a nice touch and maybe
> even a nice illusion. This is particularly a
> convinient way to gather information and opinions and
> possibly have these affect some group decision;
> however, this is not a model that will likely ever
> amount to much amongst a loose group of individuals
> that have little commitment to each other (to the
> group) -- individuals that most likely won't even
> trust any unfavorable results as there is no existance
> of one vote per one human that is trusted.
Let's analyse the Debian's community experience.
Will the GNU Herds Association be similar to The Debian community and so a
duplication of that organisation?. Reading
http://www.debian.org/devel/constitution we can see some of the vitally
differences. We will not support the 2, 3 & 6 points, only the 1.
2. Decision-making bodies and individuals
Each decision in the Project is made
by one or more of the following:
1. The Developers, by way of General
Resolution or an election;
2. The Project Leader;
3. The Technical Committee and/or
its Chairman;
6. The Project Secretary;
The trouble of using central servers
------------------------------------
> The GNU.FREE voting system is not the solution to all
> voting problems any more than any system could be. In
> particular, there is a central point where the
> potential for abuse is great. This is fine (great)
> when this central point can be watched, controlled,
> audited, etc. properly; that is, when the voters have
> near complete trust that the the system in not being
> abused or hijacked at this central point.
> As if this wasn't enough reason (for me) to lose
> confidence, pulling off a great system with great
> results would in all likelihood lead to very little
> difference since those with power and with something
> to lose would go their own way and the minority (if it
> was a different group) would just go elsewhere as
> would have happened vote or no vote.
> The whole software issue is only a small piece of the
> whole picture and needs to be designed with the wholes
> in mind.
Using the 'Peer to Peer network' paradigm instead of the 'Central servers'
each member can have a duplication of:
The web page
Before: DNS --> one IP
After: GNU Herds --> a IP set
The better IP will have greater followers.
You could search for "GNU Herds" on google
and locate ranking of GNU Herds web servers,
the servers itself, etc.
The email lists
Using the P2P idea, when a member send
a message, it must go directly to each
member email address, without the need of
an email list server. The simplest, add
to the BCC all the members email addresses.
The voting system
About the host which are used to keep the
member DB, the voting system, etc., I
think that the system must be designed to
be P2P based instead of a central server
system. All system in the P2P network get
the same member DB, votes, etc. So you can
not manipulate a system without being
noted by the others duplicated 'servers'
on the network.
Any member can connect to the voting
system and audit the voting process using
her system 'copy'.
Who fill the member data base?. The
Secretary for example. When any member
connect to the P2P system he can audit the
new information (members, votes, ...) which
his host is getting form the network.
The trouble of certificating identities
---------------------------------------
> Additionally, there has to be faith in the database
> information itself. Does John Doe Jr. really exist?
> Are maryX, peterpipper, 57-4-me, and tdozen4 really
> four separate individuals? How can I be sure?
The tendency in Internet is to use unique identities, using one or two email
addresses (home & job). Maybe, the ( "Name" <id@domain> ) is easily
associated to the same guy. So, if a member claims to be "Peter" and
collaborate with a set of FS projects, I think perhaps it could be easy to
detect duplications. Any member can inspect the member list.
Of course it is possible some guys work with a lot of Internet identities.
However I think that is not the usual way just today. So I think we could use
'Internet identities' just now, and later (one year or ...), only if it is
felt that it is needed, we could add *additional checks*, as countries does
to its citizens, giving the trust about the *additional checks* to offices
which certificate the people with photograph, fingerprint, etc.
> What opinions does anyone have of the following: to
> help draft (or find) a form of "certificate of
> identity" that can be useful in ensuring trust amongst
> the members that form a virtual (Internet) company. As
> an example, this certificate must have some personal
> information.
I think that just now the below is right to get an identity. A private-public
key is bound to a set of email addresses.
* email addresses
* Name
* List of FS contributions with references, etc.
We could add later actual offices on countries, cities in which we perform
the additional checks certificating a member-ID with:
* photograph
* fingerprint
* etc.
and checking it is unique around all the member data base. Maybe we will not
have to do any assumption.
Appendix:
Reference: http://www.debian.org/devel/join/newmaint
Series of required proceedings to become a Debian developer.
# Checklist - required steps for applicants
* Step 1: Initial Contact
* Step 2: Identity Verification
* Step 5: Evaluation and Check-in
Note from http://www.debian.org/devel/constitution
7.1. Powers
The Secretary:
Takes votes amongst the Developers,
and determines the number and
identity of Developers, ...
Others
------
> Finally, there is at least one more issue, the idea
> that the results of the vote have to be meaningful to
> someone. Sure, an opinion vote is nice, but for our
> purposes, will the vote be binding? What if the
> "leaders" or anyone with something at stake simply
> decide not to honor the results...
You can update your own servers and compete with the others. Bill, are you
back?.
P.S.
----
I think all this must be thought slowly. Lets wait several weeks up to people
show her/his opinions. Note that I have not read yet
http://www.thebell.net/papers/vote-req.pdf
GNU Herds: http://80.25.136.215/beta3/
Debian: http://www.debian.org/devel/
Reply to: