David,
If I
have the process right, I think we need a slight change.
The
scanner produces a 460, which is a collection of votes. These can use the
ShortCode and include a ballot identifier for each cast ballot. I think the next
step is what you call the tabulator - something that totals the votes for each
candidate. It is at this stage that we need something that, for each ballot
identifier, associates the candidate with the ShortCode. This means that the
ShortCode must be associated with a BallotIdentifier in the 230. TheDisplayOrder
and ExpectedConfirmationReference could also potentially depend on the
BallotIdentifier.
Do I
have this right?
Regards
Paul
-----Original Message-----
From: David RR Webber (XML)
[mailto:]
Sent: 18 July 2006 19:41
To: Paul
Spencer
Cc: eml
Subject: RE: [election-services] Ballot
counting - bridging between EML 410 and EML 230
Paul,
Just digging into the details here actually I think the EML 510 should be
totalled on ID not ShortCode most likely.
As Peter noted - you will need to associate between the raw codes coming
off the optical scanner and the EML 440 records you are creating already at a
prior step. This needs to take into account form varients etc. So
its logical to look the ShortCode to ID values up at that point. I think
Peter was hinting too you may need more than one EML 230 for each form varient
the scanner produces.
So when you get to the EML 510 totalling process - you are using ID
values that are consitent with the overall master EML 230 at that point - even
though you are - at the scanning step - using a sparse EML 230
as you have below that does not expose candidate information.
Thanks, DW
--------
Original Message --------
Subject: RE: [election-services] Ballot
counting - bridging between EML
410 and EML 230
From: "Paul Spencer"
<>
Date: Tue, July 18, 2006 10:10
am
To: "David Webber" <>
Cc: "eml"
<>
David,
I think
the ShortCode attribute of the candidate could be good for this.
The
ShortCode was intended for uses such as SMS voting, where the
candidates are
listed in order and the voter votes for number 1, number 2
etc. If we are
talking about the counting engine working in field order,
this seems to be a
similar requirement.
The 510 could then use the
ShortCode as the only identifier for the
candidate, the code just
representing the order on the ballot. Meanwhile, a
version of the 230
could be used providing just the Id and ShortCode
attributes for each
candidate.
A section of the 510 would then look
like:
<TotalVotes>
<Selection>
<CandidateIdentifier
ShortCode="01"/>
<ValidVotes>1234</ValidVotes>
</Selection>
<Selection>
<CandidateIdentifier
ShortCode="02"/>
<ValidVotes>2345</ValidVotes>
</Selection>
<Selection>
<CandidateIdentifier
ShortCode="03"/>
<ValidVotes>4321</ValidVotes>
</Selection>
</TotalVotes>
Meanwhile,
a portion of the 230 would
be:
<Contest>
<ContestIdentifier
Id="12"/>
<Candidate>
<CandidateIdentifier Id="ABC"
ShortCode="01"/>
</Candidate>
<Candidate>
<CandidateIdentifier
Id="BCD"
ShortCode="02"/>
</Candidate>
<Candidate>
<CandidateIdentifier
Id="CDE"
ShortCode="03"/>
</Candidate>
</Contest>
I
assume that this 230 would be used by the counting engine. Would this
meet
this need? Does rotation have an impact?
By the way, I did
not get your email directly. Perhaps this is related to
the new OASIS
mailing system. I have not yet had the re-registration email.
It is
possible others did not see it
either.
Regards
Paul
----- Forwarded Message
----
From: David RR Webber (XML) <>
To: Election
EML <>
Sent: Monday, 17 July,
2006 3:14:42 AM
Subject: [election-services] Ballot counting - bridging
between EML 410 and
EML 230
Team,
I'm working on an
anonymous counting mechanism for use with paper scanned
ballots
particularly - but the same could also apply to DRE votes too.
When
the ballot comes off the scanner - the scanner is just plucking
fields
and columns and X's - so it has no idea of candidates / ballots et
al.
That is a good thing.
So it merely dumps data as
<fieldID="XXX" value="1"/> where XXX is the
sequential # of the
field it scanned.
Now I want to align fieldID=023 with candidate 007
- etc, but I want to do
this anonymously - like putting beans in tins -
so the counting mechanism
has no idea which tin is
what.
Unfortunately EML 230 exposes the candidate affliation,
description, etc.
What I need is a bridging EML that maps between the
actual ballot layout -
EML 410 and the list of voting candidates /
options - anonymously - (that
will create EML 440's then from this and
the scanner output).
I can of course go ahead and create some XML for
this purpose - fieldID =
candidateID and that is all that is in that EML
- excepting header
information on the event etc for matching
purposes.
I'm just wondering if I'm missing something that exists
already?
If not - should we add this to EML 5.0
requirements?
The goal is to create a general purpose counting system
that can total EML
ballots given the EML control XML's - EML 110 - and
this other EML XML - and
then this runs thru and does all the totals by
anonymous ID value.
A final step then separately and independently
uses the final counts, and
the EML 230 to figure out who got what
votes.
Thanks,
DW
---------------------------------------------------------------------
To
unsubscribe from this mail list, you must leave the OASIS TC that
generates
this mail. You may a link to this group and all your TCs in
OASIS
at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
---------------------------------------------------------------------
To
unsubscribe from this mail list, you must leave the OASIS TC
that
generates this mail. You may a link to this group and all your
TCs in
OASIS
at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
