[Tfug] SQL database question

[David Cowell]

On Tue, 2008-03-18 at 18:47 -0700, Bexley Hall wrote:

> I wouldn't accept any "black box dump" unless you
> know what else is "under the hood".  I.e., you
> need to know how data was committed to the database
> (was it "colored" by the procedures used to store
> it??) as well as how the "results" were officially
> reported vs. how the *dump* was created.
> 
> Imagine a human agent acting as the DBMS.  You
> *tell* him you are voting "DEMOCRAT".  How do you
> *know* that this is what he is "recording"?
> 
> Likewise, regardless of what he has recorded, when
> you ask him to tabulate the results, you are relying
> on him to faithfully reproduce the data that he
> has stored (assuming that data to accurately
> reflect the data *given* to him).  How do you 
> assure yourself that if someone *else* asks him
> for "results" (in a different form -- i.e., a dump)
> that those results agree with the "reported"
> results?
> 
> Since there is no other way of auditing the accuracy
> of the *process* (from "data in" to "results out"),
> I would insist on complete transparency (i.e.,
> roll up your sleeves and *show* me there's no
> rabbit hiding up there...)
> 
What has always filled me with wonder is: "Why the devil do we make this an electronic process at all?"

Somehow, back a hundred years ago (and even more recently) we were able
to periodically muster the *manpower* to count the ballots in relatively
short order. The system was no more fallible than the present system...
and it had a paper trail that could be audited if the normal procedures
were followed.

Are we just too lazy or uncommitted, or do we just *hafta* know who won
before we go to bed at night?
>       ____________________________________________________________________________________
> Be a better friend, newshound, and 
> know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
> 
>