Re: Consensus of the OVC

From: Karl Auerbach <karl_at_cavebear_dot_com>
Date: Thu May 27 2004 - 03:40:07 CDT

By-the-way, using CDW retail prices as a metric, a 15" touch screen costs
$689 and a 17" costs $1089. Not an inconsequential difference.

> > If this project is to succeed then it needs to focus on its core
> > proposition - the paper ballot - things like "everything on one screen"
> > are nice research topics for a later day.
> >
> This is a simplistic characterization of my argument. For some ballots, it
> may not be practical to get everything on one screen. But does this mean we
> have to page through one contest at a time? What about the 76 contest
> ballot? Most of those are judicial retention. Can you imagine voters
> paging through 76 screens?

I'm sure that the judges involved don't think that their elections deserve
second tier status on the ballot.

In addition my argument is simple, not simplistic. My argument is that
unless we focus on getting things done we are going to become a "had its
chance but missed its window of opportunity" technology.

Systems are now being certified without see-all-at-once screens. We ought
to follow that lead and focus on the really valuable and unique aspect -
the paper ballot.

There will be plenty of time later to do pretty big-screen displays and
experiments.

> > We have an extremely narrow window of opportunity - once printing DRE's
> > are purchased our day is done and we become a "could have been"
> > technology. I doubt that any election official who, having risked his job
> > once by chosing DRE's with printers is going see enough value in paper
> > ballots to risk his job a second time; they won't perceive enough
> > incremental value.
> >
> I don't buy it. Voting technology has evolved over a period of decades
> (centuries, actually, but it has changed rapidly with the advent of
> computers). It will continue to evolve for decades to come.

I disagree. A typical civil servant isn't going to go out and risk
his/her job to buy a second round of voting gear right after buying a set
of nice shiny printing DRE's - there simply won't be sufficient political
or budgetary justification for it.

Technology does evolve, but the best technology does not always win -
witness Betamax vs VHS. It took a true revolution in formats, the DVD, to
finally begin to displace what everyone new to be a deficient format, VHS
tape.

The same is true for printing DRE's (which are not an oxymoron) - they
will satisify many of the concerns for voter verification. And once those
concerns are satisfied, our points about whether the paper is the primary
source of voter intent or merely an audit record are going to be heard as
esoteric minutiae and not worth the extra conversion costs.

My feeling is that if we miss the current window of opportunity then the
next chance is at least 20+ years out.

> ...It's certainly true that the sooner we can get something
> certified the better. One of the main principles we need to follow to
> streamline certification is this:
>
> STICK WITH COTS.

I see no reason for that unless it gets us to certification faster.

My feeling is that the way to express this thought is that we use industry
standard architectures that may be reified onto COTS platforms or onto
special platforms depending on the economics of the moment.

Clearly, code building goes faster on standard PC's.

Maybe even initial certification on selected standard PC's.

But commercial deployment is going to bring in many other economic factors
- not the least of which is voter and administrator subjective and
emotional perception of the solidity of the package - that I believe will
drive most counties to go for tightly integrated specialized gear with at
least a 10 year, probably longer, useful life expectency.

> We have to have an enclosure that may or may not be strictly COTS.

We have to stop thinking of "enclosure" - a box wrapped around a
collection of various odds and ends of PC versions is going to look
exactly like what it is - a kludge. It will tend to be a transportation,
storage, and maintanence nightmare and unless done with care it could
invite negative reviews from the insurance folks and fire marshalls.

I am in active discussion with multiple multi-billion dollar companies to
see whether they are willing to produce an appropriate box - including the
electronics. (All of these companies produce this kind of stuff on a
routine basis and all comprehend Linux and open source software issues.)
The feedback to date is positive but no committments yet. The target box
that I've suggested to them is a single unit, batteries inside, printer
inside, possibly the paper inside, flat touchscreen, total power draw on
the order of 75watts (maybe a bit more depending on the the printer
technology) with appropriate ports for peripherals (flash memory, CD/RW)
all behind doors that can be sealed with locks or lead-wire seals.

> For political reasons, I want to shoot for getting something certified this
> year--

In which case we ought to follow the most direct path to getting code
produced - which is to do as much copying of already certified methods as
we can legally do and invent only what we need to prove the core concept.
And to my mind that means that we use simple paginated hierarchies of
menus, just like the other guys do, and focus our energy on the paper
ballot mechanism itself.

                --karl--
==================================================================
= 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 23:18:08 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:17 CDT