Re: TED SELKER: US Election Assistance Commission--Questions

From: Edward Cherlin <edward_dot_cherlin_at_etssg_dot_com>
Date: Tue May 11 2004 - 03:49:16 CDT

On Friday 07 May 2004 15:52, charlie strauss wrote:
> Ted Selker, MIT is proposing the following voter verified
> system:
> instead of paper, record a machine generated audio readback of
> the voter's choices.

And the advantage is?

Again we will have an electronic record that in the simple case
can be altered afterward. It will take much longer to process
the audio than to scan paper. Security would require a parallel
data track with all of the security features of our bar code,
which brings us back to the need to verify both the audio and
the data. At this point the system is becoming much more
expensive than paper.

A case of paper is ten reams or 5,000 sheets. 200 cases (eight
pallettes and a fraction) is a million sheets. That much storage
space for 22 months would cost a few thousand dollars in a
commercial facility. Chump change in the overall election

> This can actually be done concurrently
> with the voting process. Selker thinks that ergonomically
> this concerrent readback may have a higher voter awareness
> that checking after placing the vote. The audio tape is
> machine readable in principle.

If you use digital audio, it becomes even more readable, and if
you add a data track alongside the audio track, it becomes
trivial. But then you have to provide machines that can read out
the data so that the voter can compare it with the audio.

> Selker noted that the concurrency of the readback is
> important. He cited one election using VVPAT where when
> voters relaized they had a bad abllot only 1 in ten chose to
> revote.

Can we get that citation?

> At first I thought, well continuous tape = Bad since it gives
> vote order. but then It occured to me that two things could
> be done to solve that issue. 1) do it just like OVC does. OVC
> shuffles the ballots AFTER the election.

That bothers me a bit. Everything needs to be written as soon as
possible to avoid loss in a crash. Writing to a database on the
hard drive and then to the CD-R at the end will have a very low
error rate. But I still think about Bill Godbout's design rule:
"Any measurable error rate is too high."

> I principle one
> could re-record the tape and randomize the order of the items
> on the serial tape AFTER the election. 2) alternatively
> rather that write to a tape, write to a write-once cd. place
> the items on this in files stored in random order on the disk.
> In both cases its clearly esseintial that the
> recodring/rerecording device is separate from the touch
> screen, and ideally is a dumb machine (to avoid machine
> collusion).
> anyhow, ignoring any dislike for Ted Selkers other opinions,
> are there problems or complications with this approach? One
> possible objection is that it would not work for deaf voters
> or votrs with no spoken language (hey dont laugh! One of the
> speakers at the EAC was worried about voters with no written
> language in new mexcico)

Only 1200 or so of the 6000+ languages in the world had ever been
written the last time I looked. Most of the unwritten languages
belong to small communities, and in those communities almost
everybody speaks some other language as well. But not everybody,
and many who do speak a majority language cannot read it.

In India ballots include candidate pictures and party icons so
that the several hundred illiterate citizens can vote. We
shouldn't have a problem with that in our system.

Edward Cherlin, Simputer Evangelist
Encore Technologies (S) Pte. Ltd.
New voices in the global conversation
= The content of this message, with the exception of any external 
= quotations under fair use, are released to the Public Domain    
Received on Mon May 31 23:17:33 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:16 CDT