with this, my bladder's empty

[Kurt Cockrum]

Rich said:
>Kurt, I'm now thoroughly confused.

But I'm not.  Does this mean I've won the argument? :)

>I never did get the specific problem with electronic voting.  You
>basically repeated that you didn't trust it, but gave no explanation.  

The problem isn't "specific", it's general, and rather abstract, and
difficult to get across to non-computer folk.  It's tough enough to
discuss among colleagues.  I wish CPSR was some help here, but they're
not.  Try explaining crytography to a lay person as an example.

Old-fashioned voting systems have the advantage of operating under
principles that are common knowledge; there are no "black-box" phenomena
with an election system.  All the parts are subject to scrutiny by
people armed with common knowledge, and so can directly verify that
things are as they appear.  They don't have to be armed with years of
experience in arcane subjects, as software people do.  All they need is
the native machinery most of us are born with, if it hasn't atrophied
because of too much cultural immersion.  It ain't rocket science.

A computer is a system that runs on components that the average
person thinks are magic, and presents the illusion of of reliability
and transparency.  I don't know how to state it any clearer than to
say that all that transparency and reliability are purely illusory.
Under the skin, the software is complex and arcane, not well understood
by the authors, much less anybody else.  Sorry, too much rocket science.

The reliability we *do* get from computers is because the parts we
see have been hammered on for many years by talented people (both good
guys and bad guys), mostly working on free software (*not* commercial).
The reliability has *emerged*, which is to say that it wan't there from
the git-go, it had to evolve.  Sure, design principles and good practice
exist, but as we have seen, this isn't enough.

Consider the filesystem on scn, for example.  Election-management software
would be easily as complex as a filesystem.  Unix filesystems are,
in general, highly reliable.  People have investments of millions of $
in the data that resides in them, so there's been plenty of incentive
to develop reliable software in that domain, and *lots* of people have
worked on that software.  If an election system could be as reliable as
a filesystem, with repair and monitoring tools and the like, including
built-in trust-enhancers, and check-and-balance-style decision-making,
under the *control* of the users, I'd have a lot more confidence in the concept.

This isn't to say it can't happen.  I think it would be a worth-while
project, if a good set of requirements and design could come up.  Software
like that, if it worked, could be as subversive against established
orders as PGP, and so, for that reason alone, is worth supporting.

But as it is, it's more like a topic for PhD candidates than a serious
policy option for SCN.  Or a candidate for a cooperative open software
development effort, a la some of the Open Software efforts going on right
now.  I have no idea if any such thing exists.

As to whether the election is conducted by mail-in ballot, or by having
to be present at a meeting, well, I really don't care.  I'm just against
ill-advised leaps on technological bandwagons.  And unless there's
home-grown talent suitable for pioneering this, I'd say, give it a miss.
Or wait for somebody to get inspired.

>You haven't solved the serious barrier the travel requirement causes.  Can
>you address that?

Yeah, I think it's pretty trivial.  Consider trying to cross the freeway,
or trying to escape from old East Germany.  Now, those are what I call
"serious barriers".  The temporary inconvenience of an annual election pales
by comparison, and I have trouble taking the concept seriously.

>> A real-world example is "just-in-time" inventory systems, which function well
>> only in perfectly harmonious conditions,
>
>RL:
>I don't see the validity in your comparison with universal voting and JIT
>inventory.

I wasn't talking about "universal voting".  I was talking about election
management and likening the instabilities of a speeded-up democratic
process to the difficulties with other hastily adopted time-saving
management technologies.  I can see where a lot of the heat is coming
from if you misread what I said that badly.

>In any case, we should never use the meeting-only voting method again.

Well, mail-in absentee ballots seem to work reasonably well, at least for
those who take the trouble to fill them out.  I live in a precinct (White
Center, "Heights" precinct) where all the registered voters (not many among
the local populace) get absentee ballots because the KC election department
thinks it's too expensive to set up polling places for that few voters.
Absentee ballots are good for overworked people with little time.  But take
note that that's not the reason the KCED is doing this.  I don't think
absentee ballots or e-mail voting would work as a tool to get apathetic
people to cast their votes because the act of voting is somehow brought
below these people's threshold of action.  If they stay away from "meat-world"
polling-places, they'll stay away from the virtual ones, too.

For election-management system horror stories and cautionary tales, take a
look at the RISKS list, or browse their archives.
Here's abridged blurb-info:
	 The RISKS Forum is a MODERATED digest.  Its Usenet equivalent is comp.risks.
	=> SUBSCRIPTIONS: PLEASE read RISKS as a newsgroup (comp.risks or equivalent) 
	 if possible and convenient for you.  Alternatively, via majordomo, 
	 SEND DIRECT E-MAIL REQUESTS to <risks-request at csl.sri.com> with one-line, 
	   SUBSCRIBE (or UNSUBSCRIBE) [with net address if different from FROM:] or
	   INFO     [for unabridged version of RISKS information]
	...
	=> ARCHIVES are available: ftp://ftp.sri.com/risks or
	 ftp ftp.sri.com<CR>login anonymous<CR>[YourNetAddress]<CR>cd risks
	   [volume-summary issues are in risks-*.00]
	   [back volumes have their own subdirectories, e.g., "cd 19" for volume 19]
	 or http://catless.ncl.ac.uk/Risks/VL.IS.html      [i.e., VoLume, ISsue].
Indulge your schadenfreude over the morning coffee!  Better than the morning
paper!  It's also pretty good for the occasional frisson up the spine.  :) :)
--kurt
 I found this quote in the blurb accompanying some Delaunay triangulation
 software from AT&T, apparently apropos an ancient language religious debate:
   I think the best way I've heard this put is  "Pascal gives you a 
   water pistol filled with distilled water.   C not only gives you 
   a loaded .357, it points it at your head as a default.   Why  do 
   you think Pascal is taught in school? And which would you rather 
   have when there was a hungry bear in the area?"    --Jim Harkins
 I'd clobber the bear with Vols 1 & 2 of the Java Class library --kurt
* * * * * * * * * * * * * *  From the Listowner  * * * * * * * * * * * *
.	To unsubscribe from this list, send a message to:
majordomo at scn.org		In the body of the message, type:
unsubscribe scn
==== Messages posted on this list are also available on the web at: ====
* * * * * * *     http://www.scn.org/volunteers/scn-l/     * * * * * * *