Re: [OVC-demo-team] Random ballot generator

From: Alan Dechert <alan_at_openvotingconsortium_dot_org>
Date: Sat Mar 06 2004 - 03:13:24 CST

Thanks for taking the initiative, David. I'll try it out.

> I have gone ahead and written a random ballot generator. It's usage is:
>
> $ ./random_ballots.py <<num>>
>
Like it.

> You can generate as many sample ballots as you like, all in the working
> directory (as long as you want fewer than 10000, since that's all the
> ballot-id's there are to go around). The ballot-ids chosen are unique
> and random.
>
> These ballots match the sample that Jan posted at:
>
> http://gnosis.python-hosting.com/voting-project/OVC-Demo/0182.html
>
> Or directly:
>
> http://gnosis.python-hosting.com/voting-project/OVC-Demo/att-0182/v-
> 20041104-US-CA-Santa_Clara_County-2216-4310-213.xml
>
> Moreover, I utilized Jan's contests.py file to configure the candidates:
>
> http://cvs.sourceforge.net/viewcvs.py/*checkout*/evm2003/src/evm2003/
> data/contests.py
>
> A few notes on what I did:
>
> * I wasn't sure about the "serial" attribute (which Alan recently
> described as the CD number). I just hardcoded the value Jan provided
> in his example. Should this be a switch? Likewise, all the other
> constants like date, precinct, county, etc. are just defined at top. I
> could pull them out to a configuration file or the like, if needed.
>
I want to have a number on the label.

Actually, this is a bit mysterious to me: I noticed that when I burn a CD I
get a number like xxxxxx_xxxx for a label that shows up when I "explore" the
disk. I don't know how this gets generated. I assume we can control what
gets written there. I don't know if there is another place a "serial
number" can be written. But, anyway, the idea is that the CD should have a
numbered burned on it so the software can match the CD with the voting
machine.

In other words, I am thinking that with the production system, each CD will
have a unique serial number that will be registered on the admin PC before
voting begins.

> * I do not zero-pad the "number" attribute (ballot-id) in either file
> names or root attribute. Should I?
>
Yes. We need 4 digits.

> * I -do- enable "no preference" votes for "pick-one" contests, but I do
> not enable random write-ins. Is this important to do for the demo?
>
It's not real important. It might be nice to have some write-ins to
illustrate the difference between the barcode generated XML files and the
voting machine generated XML files.

> * However, for the multi-pick contests, I enable write-ins, mostly
> because County Commissioner lets you rank up to eight, but there are
> only seven names listed (so I need a write-in if 8 ranks are used).
>
okay

> Side note: how would a voter write-in multiple names in multi-pick
> contests?
>
I didn't accommodate that in the ballot design. That's a special feature of
the demo. Some might call it a bug.

Actually, whether or not you have any space at all for write-ins depends.
In some states, you only have a space for a write-in if there are candidates
that have registered as write-ins. I guess you could say that for the mock
ballot we have assumed that there is one write-in candidate that has
registered in these multi-pick contests.

The multi-pick contests (or N of M and Ranked Preference) were included to
show that we can accommodate different scoring methods in these cases. We
may or may not have time to show these different scoring methods for the
March demo... maybe April. Chris Schaefer did some work on this but right
now I consider this extra credit for the March demo.

> * If no preference is expressed in multi-pick contests, that is
> reflected in the absence of <selection> elements, not by the explicit
> phrase "No preference indicated" in a selection. This makes more sense
> to me, but I can change it if someone tells me this is wrong.
>
I don't recall if we decided anything on that. What does Jan's app do in
this case?

> * Contrary to Alan's request, I HAVE NOT weighted the votes against
> certain candidates (mostly because the random picks are more generic
> than that...
>
I see.

> but also I hardly like the idea of calling Marx a "bad
> guy" :-)).
>
I was afraid of that. And if Lenin/Marx "wins the election" we get labelled
as communists.

> Still, a demo ballot set can be pruned to get desired
> outcomes, if needed.
>
Right. We should consider the outcomes. People will notice if all the
Republicans lose.... or all the Democrats lose.

Thanks, again, David.

Alan D.
==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Thu Apr 1 02:40:24 2004

This archive was generated by hypermail 2.1.8 : Thu Apr 01 2004 - 02:40:36 CST