[E-voting] Today's Irish Times

Michael McMahon michael at hexmedia.com
Thu Jul 13 11:55:04 IST 2006


>
> On Thu, Jul 13, 2006 at 09:59:58AM +0100, A.J.Delaney at brighton.ac.uk <http://lists.stdlib.net/mailman/listinfo/e-voting> wrote:
> >/ For my money you can't find a more widely respected software engineer
> />/ than Prof. David Parnas, who I believe was on the CEV.
> /
> Prof. David Parnas made a submission to the CEV and also did some work
> for the commission very early on. He is not a member of the commission
> on electronic voting, I'm not certain if his work for the commission has
> been ongoing or not. Last time I met him, I got the impression that it
> wasn't.
>
> Personally, I'm not convinced of the quality of his work in this field,
> and as a result of talking to him about E-voting (both in person and via
> e-mail) can only conclude that he (at least at time) was woefully wrong
> on the issue, and wanted to frame it as purely a software reliability
> issue.
>
>   
I agree with you. I wasn't impressed with his (brief) submission to the 
CEV.
He quite cleverly trod a middle line between opponents who allegedly cite
"the laws of computers" (something I had never heard used)
and supporters whose arguments he partially rejects. It was a regrettable
mispresentation of the arguments against.

The approach taken by the CEV seems to be along the lines of his experience
with mission-critical systems, but he didn't seem to appreciate what the 
big difference
is, between e-voting and typical mission-critical systems.

I've only had a chance to scan the CEV2 report, but the chapters seem to 
vary
in quality significantly, presumably since they were written by 
different consultants.
Some of the analysis is good, but the recommendations look to me rather 
vague
and impractical, and not really justified by the analysis. Same goes for 
the overall
conclusions. Most disappointingly, the report does not seem to address VVAT
in any significant way; neither a case for it, nor against. It seems 
VVAT remains the
elephant in the corner of the room, who everyone is delicately stepping 
around.

Flawed as it is, the report does not just recommend changing the IES. It 
is also
recommending changes to the voting machines. It is completely 
unjustifiable to
start a costly process of re-designing the IES software without doing a 
cost/benefit
analysis of *everything* the CEV is looking for. Never mind that the end 
result won't
be acceptable anyway, but they must know that by now.

Michael



More information about the E-voting mailing list