RE: [election-services] Re: EML AUDIT PROGRESS

["Zelechoski, Peter" <pzelechoski@essvote.com>]

Neal -

So, let's say you have a ballot with 2 contests.  One contest is a vote for one.  The other is a vote for 2.

Ballot 1 has one vote in contest 1 and one vote in contest 2
Ballot 2 has one vote in contest 1 and 2 votes in contest 2
Ballot 3 has no votes in contest 1 and 2 votes in contest 2
Ballot 4 has no votes in contest 1 and 1 vote in contest 2
Ballot 5 has no votes in contest 1 and no votes in contest 2
Ballot 6 has 2 votes in contest 1 and no votes in contest 2
Ballot 7 has 2 votes in contest 1 and one vote in contest 2
Ballot 8 has 2 votes in contest 1 and 2 votes in contest 2
Ballot 9 has 2 votes in contest 1 and 3 votes in contest 2

That should cover the basic possibilities.

So, we should have something that says we had:
9 ballots cast, 7 ballots cast with under-votes, 4 ballots cast with over-votes, 1 ballot cast fully voted
2 valid votes, 3 under-votes, and 4 over-votes in contest 1
9 valid votes, 7 under-votes, and 1 over-vote in contest 2

But there are many places where they would say a blank ballot is a special case or a "protest vote" which means Ballot 5 should be counted differently.

There are some places where only the partial vote is considered an under-vote so ballots 3,4,and 5 are not under-votes for contest 1; also, ballot 6 is not an under-vote for contest 2.

I think the simple set of numbers is the most prevalent in the US and what most people would think if they didn't know better.  So, maybe we should focus on that first and worry about the local exceptions later.

Votes is something that is at the contest level (and then at the candidate level).

Ballots needs to be above that.  I see where you brought it down to the Contest-Ballot level; is there a reason to do this?

I see you have a Precincts Reporting CountMetric.  I look at precincts as a type of reporting unit; other types I think of are Central Site, Central Site (Election Day), Central Site (Absentee), other ...

I found the new requirement for the StatusDetails item to be cumbersome -- I would much prefer it to be a min occurs of zero.

- Peter 

-----Original Message-----
From: Neal McBurnett [mailto:nealmcb@gmail.com] 
Sent: 2010-11-22 6:22 PM
To: Joseph Lorenzo Hall
Cc: John Borras; EML TC
Subject: Re: [election-services] Re: EML AUDIT PROGRESS

Our audit in Boulder has gone very nicely.  My ElectionAudits software
now supports Kaplan-Markov audits with strict error bounds.  I've also
collected a bunch of ideas on audit reporting from others in the US
and sent them on to Joe.

I'm glad to see more concrete proposals for EML and P1622 from NIST
and a timeline.

The next P1622 meeting will probably be the week of Dec 13th.

Can we schedule an election-services meeting the week before
that, and give Joe and I more concrete incentive to get something
together for the committee?

Note that I'm still looking for some examples of how to label blank
votes and overvotes in a 510 report via the current EML spec, which is
confusing to me in that regard.  E.g. is there a way to use the
BallotStatus property in a 510?  If so, if someone could just adjust
my current example to properly label them I'd apprecate it:

 http://bcn.boulder.co.us/~neal/electionaudits/eml510-example.xml

As I indicated in my earlier comments, I expect we still need to find
a precise way to report the number of "contest ballots" in a
vote-for-more-than-one contest, but perhaps the exisitng scheme can
work for "vote for 1" contests.

Cheers,

Neal McBurnett                 http://neal.mcburnett.org/

On Mon, Nov 22, 2010 at 06:14:01PM -0500, Joseph Lorenzo Hall wrote:
> On Mon, Nov 22, 2010 at 4:12 AM, John Borras <johnaborras@yahoo.co.uk> wrote:
> > Hi Guys
> >
> > Now that the USA elections are out of the way, have you been able to make
> > any progress on the audit requirements for EML v6.1 please?
> 
> I have been slammed with research this entire month and it doesn't
> look like it will get better soon.  I know Neal has also been wrapped
> up in CO post-election auditing work, although he may have more time
> going forward than I do.  I'll touch base with Neal soon. best, Joe
> 
> > IEEE/P1622 and NIST are planning to hold a workshop in February to
> > essentially decide which standard to pursue for their solution for the
> > UOCAVA project and we will be invited to demonstrate how EML could fit the
> > bill.  NIST want a P1622 draft standard to be available by June 2011 for use
> > in the next presidential elections so their work is now gaining real
> > momentum and I would want EML to be there in the forefront of it all.
> >
> >
> >
> > Do we need to have a conference call to discuss progress, issues, next steps
> > etc?
> >
> >
> >
> > Regards
> >
> > John

---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  Follow this link to all your TCs in OASIS at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php