🚀 go-pugleaf

6 July 1992


Colorado Daily
839 Pearl Street
Boulder, Colorado

Boxed Op-Ed piece.  (_Not_ a Letter-to-the-Editor).


In a recent article Evan Ravitz pointed to a "successful election" by
telephone in Canada.  The truth is quite different.

First, this "election" was actually an extended Liberal party caucus
open _only_ to members of the Liberal party.  It was not open to the
general public.  This has three major implications:

 o There will be an artifically high 'voter turnout' since people
   who bother to affiliate with a party are generally more politically
   active than independents,

 o None of the constitutional safeguards regarding public elections
   are required.  A political party is a private organization which
   can choose candidates in any manner they choose (in the U.S., both
   the Democrats and Republicans water down the popular vote with
   "automatic delegates" including member of Congress, Governors,
   state party chairmen, etc.)  A state has constitutional restrictions
   on its elections, and

 o There were only 7289 registered Liberal party members in Nova
   Scotia.  The city of Boulder, in contrast, had 52,051 registered
   voters as of last week.
   
   In addition, the _proportion_ of eligible voters is greater in
   Boulder since the Canadian experiment was limited to members of
   a single political party.  This has a serious impact on the
   physical infrastructure such as phone lines.  (The fastest
   way to bring down the phone system is to take approx. 3% of
   all phones off the hook at the same time.  Radio stations
   offering large prizes have accidently done this several
   times).

Second, the results of this "election" were thown out due to major
compromises in the system.  The vote was supposed to begin at 12:30
and last 90 minutes.  What actually happened:

12:30  Voting begins.

12:35  Many voters report they did not get the "thank you" message
       from the candidate which indicates the vote was recorded.
       There is confusion whether the vote was actually recorded.

1:00   Voting suspended.  All results thrown out at 1:40.

2:30   Voting starts over.  Some people get "thank you" on first
       try, others try 20-30 times unsuccessfully.

3:00   Voting extended until 4:00.  Many voters complain that their
       PIN has been rejected.

4:00   Voting extended until 5:00.  Less than have of the conventioneers
       have been able to vote.

5:00   Voting is extended until 6:30.  Remember, a single round was
       only supposed to take 90 minutes, but this has round has alreadly
       lasted nearly twice that.

5:30   Reports arrive that members have been able to vote twice.

6:30   The convention is called off.

At last report, the Liberal Party has decided to try once again, but
will have a paper-ballot system on hand.  Meanwhile, the telephone
company was required to post a 350,000$Cdn performance bond.


Could this have been nothing more than poor a implementation?
Is "voting-by-phone" still a reasonable goal?

The answer is no:  "voting-by-phone" has serious fundamental flaws.
Under the proposed system you _must_ identify yourself to the state
when casting a ballot; there is no way for you to prevent the state
(or others monitoring the phone line) from recording your vote.

Mr Ravitz suggests that this would be made illegal.  That sounds
like a good idea, until you consider who is responsible for enforcing
this law.

Unaddressed is the concern that votes could be coerced.  Under the
present system it is impossible for someone to hold a gun to your
head while you vote, or for them to know how you voted anyway. 
Under the "voting-by-phone" system, a person could be coerced
without anyone else's knowledge.

A number of people, including myself, have investigated secure
electronic ballots.  It appears technically possible, but requires
extensive use of advanced cryptographic techniques presently covered
by patents.  It will never be possible to simply punch in a 8-digit
PIN to achieve secure electronic ballots.


Bear Giles
Boulder, Colorado

//The author is a professional software engineer at NOAA with
8 years experience.  Among his interests is cryptography and
its applications.


cc:
	Boulder County Clerk
	Daily Camera
	USENET group: boulder.general