The Philippines 2010 general election: Audits
The field test held on 27 January addressed all aspects of the automated election system: voting, transmission, counting and consolidation/canvassing. It proved the accuracy, security and reliability of the automated voting and counting machines.
This test proved successful. All results were electronically transmitted to the municipal, provincial, national and central consolidation centers immediately after closure of the polls, using two transmission methods: GSM/GPRS and BGAN.
The main test involved ten polling centers in urban and rural areas in metro Manila and in the provinces, five municipal canvassing centers, one district canvassing center, three provincial canvassing centers, two national canvassing centers, one central canvassing center and one broadcasting center.
There were ten optical scan machines, plus ten for contingency, 13 canvassing units (laptop and printer) and 13 backup canvassing units (laptop and printer) deployed in the field.
A second field test was held on 29 January, solely to show the media how field tests are conducted. The scope of this test was smaller (about half) than the 27 January one.
The mock election was carried out exactly like a regular election. Voters had to check in, received a ballot, marked the ballot, scanned the ballot and deposited it in the ballot box.
The final version of the election software was used and covered actual voting, counting, transmission of precinct results, and consolidation of results from all canvassing levels. All the systems to be used on Election Day were tested and used by actual voters, actual optical scanners operators, and real CCS operators.
This event involved nine polling centers, five municipal canvassing centers, one district canvassing center, two national canvassing centers, one central canvassing center and one broadcasting center.
Equipment used included nine optical scanners, plus nine for contingency, ten canvassing units (laptop and printer) plus ten for backup. Field personnel included nine technicians for the optical scanners plus nine backup, six CCS technicians (plus six backup), and ten NSC agents.
This audit was done to obtain systems acceptance by COMELEC. It was useful to test the system’s accuracy, reliability and integrity. The test validated the performance of the following items:
- Smartmatic’s Election Management System (EMS)
- Vote counting using the SAES-1800 optical scanners
- Smartmatic’s Real-Time Electoral Information System (REIS) for consolidation and canvassing of election results from optical scanners transmitted through public telecommunication networks
The PreLAT testing comprised a full, functional test of all system components, including:
- Hardware (optical scanners, modems, security keys, ballots, printers, and CCS stations)
- Software (optical scanners firmware, modem firmware and canvassing application); and the
- Telecommunications network
Each unit prepared for dispatch was started and zeroed, after which the ballot configurations for precincts were fed.
Upon completion the Election Return (ER) was printed (zero votes) and results transmitted to the same canvassing sever to be used on Election Day, using the same public network to be used on Election Day.
Once the test was complete, the units were re-zeroed (every optical scanner immediately, the CCS units only when all optical scanners of that region have completed the test), and then packed and stored for dispatch.
Source code review
Source code of the EMS, optical scanners and canvassing applications were made available for review from 10 February, 2010, in accordance with Philippines law (Republic Act 9369).
As is standard practice, Smartmatic helped the Commission set up a controlled environment ensuring all necessary safeguards for conducting the review – and made available assistance from its senior development staff.
Sealing and testing day
Testing and sealing activities were conducted 7–3 days before Election Day, when all the optical scanners units were deployed to the polling places. For the testing, the Board of Election Inspectors (BEIs) performed the following:
- Optical scanner diagnostics
- Opened the election
- Voted using ten test ballots
- Closed the election
- Printed the ERs (Election Returns)
- Election Returns (where the number of votes per candidate is printed) were then be compared with a manual tally to be conducted by the BEIs. This activity had the following objectives:
- Make sure that the equipment was working correctly post-deployment
- Show that the results from an electronic count and a manual count were the same