[Theforum] Changes to the VotingReqs document

.jeff jeff at members.evolt.org
Mon Feb 4 16:25:37 CST 2002


elfur,

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> From: Elfur Logadottir
>
> | unfortunately, total anonymity means there's no way
> | for me to guarantee that you only vote once.
> | presentation layer anonymity would mean no one but
> | those with db access could see who voted on what (and
> | even then it would require some work).  i think in the
> | end though there's no real way to do it and provide
> | complete anonymity.
>
> not giving up :)
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

too bad.  the rash went away much easier.  ;p

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> i'm assuming that the db spec says you will store the
> userID along with the vote result and that way you can
> look in the results for certain ID and change that
> record whenever changes ones mind during voting period.
> please correct me if i'm wrong.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

yes, a userid will be stored with the actual vote from that user.  i'm not
sure i like the idea of changing a vote once a vote is submitted, but if the
case for it is strong enough then i suppose the application should support
that.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> what about a slightly different approach when (and only
> when) we want total anonymity (and i do see the need for
> total anonymity, rarely, but i still do).
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

personally, i'd rather stay a *long* way away from exceptions.  in other
words, if this is a feature that the application must support, then *all*
votes/polls would use the database structure created to support this
feature.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> What about storing the userID in a different place when
> we want anonymous voting. even in its own table named
> ANONYMOUS_VOTING or something ??  that way you write the
> vote result to two places ... the result table, without
> the userID and the anonymous table (only the userID and
> the questionID). this would call for a flag of the type
> of voting, but we're doing that anyways for the voting
> vs. polling.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

so what you're proposing is having two tables.  one of the tables stores the
voteid, questionid, and optionvalueid.  another table stores the voteid,
questionid, and userid.  this could create the anonymity because there
wouldn't be any connection between the what their selection was and who they
are.  however, that completely removes the ability to allow a user to change
their vote while it's still open, which i'm actually fine with.  however, if
we're going to allow this total anonymity, we're going to have to decide if
it's more important than allowing people to change their votes.  i don't
think it's prudent to try to support both and create all kinds of exceptions
in the database architecture.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> this way you can have the 'normal' design with the
> exception that we do allow for totally anonymous voting
> when the need rises.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

no exceptions.  pick one or the other.

thanks,

.jeff

http://evolt.org/
jeff at members.evolt.org
http://members.evolt.org/jeff/





More information about the theforum mailing list