Thursday, June 25, 2009

The Last Time Comelec Tried Automation...and the Next

At the IBLOG5 Bloggers Conference held at Malcolm Theatre of the U.P. College of Law last month, I was seated beside the Comelec's Spokesman, James Jimenez who has been a blogger for many years and is the sort of government official who takes public abuse with apparent good humour and equanimity. I suppose that is a necessary qualification to work at Comelec!

At the time, there were still three live bidders under consideration by Comelec to automate the 2010 elections (AMA, Israeli firm Gilat, and Smartmatic). But there was also a lot of talk that there would be a failure of the bidding process and 2010 would have to be conducted manually. But James gave an upbeat presentation of the Comelec's automation plans and confidently predicted there would be a clear winner whose system would be acceptable "even to people like you"--whereupon he cast a friendly but chiding elbow and evil eye at me. Sitting on the front row of IBlog5, James Jimenez reminded me about the last time that Comelec tried to automate the elections. That was under Ben Abalos in 2003 with his ill-fated MegaPacific Automated Counting Machines and the billion peso bidding fiasco that was so severely dealt with by the Supreme Court in its en banc decision, penned by Chief Justice Artemio V. Panganiban: Information Technology Foundation v. Comelec [G.R. No. 159139. January 13, 2004]

My favorite passage of this SCoRP decision striking down Ben Abalos' scandal tainted automation project of 2003 is the following paragraph that happens to contain Footnote 54:

J. Panganiban: But there is still another gut-level reason why the approach taken by Comelec is reprehensible. It rides on the perilous assumption that nothing would go wrong; and that, come election day, the Commission and the supplier would have developed, adjusted and “re-programmed” the software to the point where the automated system could function as envisioned. But what if such optimistic projection does not materialize? What if, despite all their herculean efforts, the software now being hurriedly developed and tested for the automated system performs dismally and inaccurately or, worse, is hacked and/or manipulated?[54] What then will we do with all the machines and defective software already paid for in the amount of P849 million of our tax money? Even more important, what will happen to our country in case of failure of the automation?

And here indeed is Footnote 54 of the above SCoRP decision:

[54] In the December 15, 2003 issue of the Philippine Daily Inquirer is an item titled “Digital ‘dagdag-bawas’: a nonpartisan issue” by Dean Jorge Bocobo, from which the following passages appear:

“The Commission on Elections will use automated counting machines to tally paper ballots in the May elections, and a telecommunications network to transmit the results to headquarters, along with CDs of the data. Yet, with only five months to go, the application software packages for that crucial democratic exercise--several hundred thousand lines of obscure and opaque code--has not yet even been delivered in its final form, Comelec Chairman Benjamin Abalos admitted last week.

“My jaw dropped in amazement. Having built software for General Electric Co.'s medical systems business and military aircraft engines division (in another lifetime), I have learned the hard and painful way that 90 percent of unintended fatal problems with complex software lies in the last 10 percent of the code produced. From experience, I can assure you now with metaphysical certainty that not even the people furiously writing that software know whether it will actually work as intended on May 10, much less guarantee it. Simply put, the proposed software-hardware combination has neither been tested completely nor verified to comply with specifications.”
It's flattering to be quoted by SCoRP in an historic decision, but I feel absolutely no regrets whatsoever that Abalos' Automated Counting Machines were TKO'd by the Supreme Court in 2004, nor any remorse that I had played some very small role in stopping it. But, of course, after the 2004 national elections, I did wonder if the Supreme Court had not in fact been led by the nose to scuttle a system that the likes of Virgilio Garcillano, Bedang Bedol and their ilk probably did not want in place at that point in history. I doubt that we shall ever know...But fast forward to the present and James Jimenez's prediction or promise that 2010 would be an automated election.

I was skeptical at the time, sitting there with him at IBlog5, but now, lo and behold, the Poll Automation Law shepherded through Congress by Sen. Dick Gordon has apparently resulted in Comelec selecting a system and service provider to undertake the historic first-ever automated Philippine national elections in 2010. In its Notice of Award to the winning consortium bidder, Total Information Management Corp. plus Smartmatic, Inc., the Commission said:

You are hereby notified that pursuant to En Banc Resolution No. 8608 dated June 09, 2009, the Commission AWARDS to your joint venture the contract for the procurement of counting machines, including the supply of ballot paper, electronic transmission services using public telecommunications networks, training, technical support, warehousing, deployment, installation, pull-out, systems integration and overall project management, for the Automation of the Counting, Transmission and Canvassing of Votes for the May 10, 2010 Synchronized National and Local Elections, particularly:

...Particularly 7,191,484,739.47 pesos worth of goods and services.

Something James told me at IBlog5, which really stuck to my mind, was a conceptual distinction he made to me about automated systems being TAMPER-PROOF or being TAMPER-EVIDENT.

Speaking to ABSCBN News this past week, Comelec Spokesman James Jimenez mentions the fact that the Smartmatic system of 82,000 plus counting machines will be transmitting RAW PRECINCT DATA to a Comelec Database that could theoretically provide a strong measure of transparency would go a long way to securing the election. The idea here is that the "addends" that will eventually be canvassed and summed up for final vote tallies will be open to the public.

But the details of this open public database are unknown:

(1) Will it be online such that an outfit like Namfrel or a Bloggers Consortium, or some large nonpartisan citizen's group, could indeed organize an independent canvass using the exact same data as the Comelec and Congress--at least for the President, Vice President and Senators?

(2) Will it have the bandwidth ("big iron") to service the likely flood of public queries and organized efforts at auditing the results? Or will the thing bog down on Election Day and the Comelec will sheepishly apologize because they put some poopy laptop in to service the gigabit per second incoming query data?

(3) What would happen in the case of massive discrepancies? Can an independent group get special Comelec certification to help specify and build this major aspect of transparency? There are after all still 4 billion pesos left in the automation budget!

People who are interested in this aspect of the 2010 Automation Election need to get organized around the effort RIGHT NOW! The election is less than a year away.

No comments: