Re: Yes to Publicity! Not to Anonimit! Was: Re: GNOME Foundation Annual Elections - proposal

[Danilo Segan <dsegan gmx net>]

недељa, 14. септембар 2003. 07:44:50 CEST — Andreas J. Guelzow написа:
> The current method may not have any great advantages for the
> Foundation, but hte proposed method appears to have significant  
> dasadvantages.

Not any disadvantage that current method doesn't have.

> Well, there have been claims that independent counts are possible but
> since the provided data can not be checked those are indeed just
> claims. By having an anonymous vote we put all our trust into those  
> that administrate the election, effectively those that are currently  
> `in power'. While we may verify that our `code' is correctly  
> associated with our vote there does not seem to be any mechanism to  
> guarantee that all listed votes are indeed cast by valid voters:

If we provide a list of all who voted, in the discussions like the one  
you mention below, it would also come to surface if someone didn't  
vote, but was included as if (s)he did.

> say the foundation has 400 members, 50 of them in fact cast a ballot
> but the list of vote lists 100 codes and corresponding votes. Those
> interested in the election (and having cast their vote) may check the
> result and verify their own ballot but that are only 50. the  
> remaining members are unlikely to check since they have already shown  
> not to be interested in that election. Even if they do, each has a  
> 75% chance that it is correctly recorded that they did not cast a  
> vote. Since those that are interested in the election have no  
> opportunity to comment to any non-voter: "I see you voted for X, but  
> you always said he was a bad ..."
> it is unlikely that any vote tampering would be detected. (Note that
> to fix most votes you only have to add a few votes, so realistically
> probably a list of 65 votes would suffice and only 15 of the 350
> non-voters would be in the position to recognize the fraud,)

The same is with the current system. It is fairly unlikely that frauds  
will be recognized with such a "non-formal" way of discussions about  
the choices one made.

At the same time, along the same lines, one could list all those who  
voted anonimously, so if I see you in the list of the ones who voted,  
in a friendly discussion I might ask you (the same as your example  
above): "how many of your 'chosen' ones were actually elected?"

This would not intrude into your privacy, but if you didn't vote (which  
would mean fraud), we've got the same non-formal mechanism of detecting  
fraud in such a new system, like the mentioned one in current system  
(discussions with friends about votes they cast).


So, I think that on this particular issue of trust, there's nothing  
different in current and proposed system, but there surely is room for  
improvement (though I cannot readily suggest what that improvement  
should be). So if you're to go as far as to claim the proposed system  
as untrustworthy (because only non-formal ways are used to detect  
fraud), the same is with the current system.


Cheers,
Danilo