Committed to PEOPLE'S RIGHT TO KNOW
Vol. 5 Num 1055 Mon. May 21, 2007  
   
Point-Counterpoint


The matador


Interestingly, the Chief Election Commissioner (CEC) in his recent interview compared our upcoming parliamentary election to a Spanish bullfight. I presume that the CEC is playing the role of, el matador, or a bull-fighter fighting this deadly game, where he has to be fully prepared before he steps into the arena to face the bull and he must show good performance and avoid deadly mistakes to win the fight.

Unfortunately our matador is not prepared even after four months to enter into the arena as he is very careful not to make any mistake that can leave him scarred for life. But I am glad that Election Commission (EC) finally realized that traditional method of door-to-door voter registration is not feasible as they have to carry around all the digital data collection equipment like laptops, finger-print scanners, and cameras.

However, political parties believe that voter listing from camps will not work and urged EC to reconsider the collection of data from door to door. Surely, the camp-based voter listing will end up administratively disenfranchise many people, especially conservative rural women and elderly from their right to vote due to distance and long registration process.

Similarly, door-to-door voter registration may leave out the ever increasing mobile and homeless population in major cities like Dhaka and Chittagong. Although, EC believes that with proper campaign through political parties, NGOs and other mediums it would be possible to bring people to camps to register but it is yet to be proven.

There are clear advantages and disadvantages of both door-to-door and camp based voter registration. Thus this unique situation in the country demands a unique and pragmatic solution where we can use a hybrid solution using all the option of door-to-door and camp-based registry using both manual and digital method.

After going through several proposals from BASIS, BUET, Bangladesh army, and other organizations, EC now came to a decision to deploy 12,000 teams equipped with laptops, document scanners, finger-print scanners and web-cams to registration stations all across the country. Turnaround time for voter registration has been estimated at 20 to 25 minutes per voter.

If you do the math for 90 million people to be registered by 12,000 units where each registration process takes 25 minutes, it will take more than 20 months just to register the voters. If you include procurement time, training, printing of lists, logistics and holidays this time frame can reach up to 28 months to complete the entire project.

It seems that EC is more concerned about the hardware and technology that will be used in the registration process rather than finding an optimized process and right technology for registration. What we need here is to see an optimized voter registration process and methodology where we can bring the turn around time for each voter to complete registration in 5 to 10 minutes using least number of equipment and manpower possible.

Before suggesting a workable solution for voter registration, I would like to make following points that needs to be considered in light of current decision by EC and demand by the political parties:

Digital picture: Using web-cam is not at all desirable as the whole purpose of the web-cam is to do video conferencing and not take quality pictures. Instead good quality digital cameras with flash and tripod are required for the job. There should be a methodology of taking picture with white background and proper lighting condition so that picture taken can be clearly printed and later be used for identifying the voters.

Fingerprint: It is not clear why EC would want to take fingerprint scans of the voters if these fingerprints will not be crossed matched for duplicates. In order to cross match the fingers you need AFIS, an expensive solution for fingerprint verification and matching. If EC plans to match 1 to many or 1:n fingerprints, what is the number for "n"? If you are matching fingerprint for each voter with the rest of the 90 million voters then the time required for finding match will be extremely long and may not be feasible.

Data entry resources: Large number of data entry resources skilled in Bangla typing will be very challenging to hire and to deploy them all over Bangladesh for only 18 months.

Database and security: As we are relying on digital data for the registration, we need to ensure proper database security and disaster recovery planning. How will the data be transferred to the national data center from the camp location? How can the database be secured so that it cannot be manipulated or corrupted.

Turn around time: 20 to 25 minutes for each registration to complete is far too long and many voters, specially peasants and laborers may not have the time or patience to wait that long just for the registration purpose. It should be made simple and brought down to 5 to 10 minutes.

Inclusiveness: Is this new registration process inclusive, that is, how accessible is it to all groups and categories of eligible citizens. Are we not administratively excluding voters such as women and elderly by setting up camps for registration where they have travel long distance from their home and how will laborers invest in times that they don't have.

Keeping above points in mind, we can think afresh to look for a hybrid solution that will allow registration of all types of voters such as rural people, mobile population of the cities, homeless and even handicapped people. EC can look into following suggestion to incorporate them into their planning:

  • Instead of deploying 12,000 teams armed with laptops and other gadgets to camps or door-to-door, EC should consider deploying 12,000 teams with pre-printed unique application forms with bar-code, digital camera with flash, tripod and white backdrop.
  • After collecting voter information and photographs from the field, each EC team will return to Union or Thana data centre to upload the pictures taken during that day and scan the corresponding applications to the computer. Later a customized computer application can search and integrate the picture with the scanned application through background matchmaking process using the bar code. At this point data entry can be done directly from the scanned form or send it to the central head quarter for outsourcing data entry service.
  • Under this process voters do not have to wait around for data entry operators to enter their data into computer and this process is done in parallel while other mobile teams are busy collecting data in phases.
  • Here process of data gathering, data entry, validation and printing is broken down and each process can be staggered throughout the country and can be done in parallel to save time.
  • Once the data reaches the central database in the national datacentre, it will go for printing voter list with pictures through an automated digital press.

The biggest challenge for voter registration in 2007 will be inclusion of voters' photographs during the registration process. Only systematic and clearly identified processes utilizing modern technology can ensure a proper voter registration.

The CEC has rightly compared this election to the Spanish bullfight and he cannot afford to make any mistake nor does he have any time for experiments to win this fight. He has to make sure that the processes and technology chosen for the voter registration is practical and fool-proof.

Mahboob ur Rahman is an IT professional.
Picture