EVERYBODY knows the failure of the then Election Commission in its attempt to introduce voter ID card in 1995. Not only was the money wasted, but its failure is also often cited as a negative example. And raising that excuse, a number of large IT projects were handed over to the foreigners, instead of local experts, for execution. 1995 to 2007, more than a decade has passed, technology has advanced a lot. One might think that it is very easy job to introduce voter ID cards by merely pressing some keys on a key-pad. Is it really so?

"Voter ID card" together with "national ID card" have again become the catchwords these days. In the last one month, we read about more or less 100 proposals in the newspapers and the Internet on the development of ID card. Among these proposals the attractive ones were those that claimed that their proposals could be implemented within a short period of time and a relatively small budget, although most of them do not have experience of developing large IT systems.

In 2004, the Ministry of Home Affairs attempted to develop "Machine Readable Passport and national ID card." A ten-member committee was formed with the additional secretary (who is now a full secretary) of the ministry as convener. I along with my colleagues, Professor Abul Kashem and Professor Jobair Bin Alam from BUET, were members of that committee. Apart from us, the other members were representatives from the Foreign Ministry, Passport Directorate, BTTB, DGFI and NSI.

We visited Canada, Pakistan, Maldives, Mexico, Malaysia and Singapore to study their systems and to gather experience. Since the background and development history are different in different countries, it is not possible to follow the system of any one country in particular in developing ID card or passport system for our country. Based on the experiences of these countries, and considering our socio-economic condition and status of ICT, we designed a system most suitable for our country

and, accordingly, we submitted a report on that. On the basis of the recommendations of that report, it was decided by the highest executive body of the then government that after procuring the required hardware and appointing the necessary manpower, the issuance of national ID card and machine readable passport would commence from June, 2005.

In the report we have shown that six years would be required in order to include the whole 150 million population of the country in the national database. To run the system the Ministry of Establishment approved the manpower set-up for the proposed "National Registration Department (NRD)." The NRD, through its offices at the district and upazila/thana level, will create a national database containing the relevant information of all nationals. The data required by any agency like the Election Commission, passport directorate, police department, banking sector, NBR can be obtained from the National Registration Department. The fund required for procuring hardware, for the development and procurement of software and for the manpower was estimated to be 1400 crore taka.

We recommended implementation of the project by the local people after relevant training.

We had several reasons why the job should be done locally. First of all, we found that both Malaysia and Pakistan wasted huge amounts of money and time on their first attempts of implementing the project by the foreigners. Later, when they engaged local experts, they were successful. Secondly, the maintenance and running of foreigner-implemented projects is always very difficult when they leave the project. ID card and passport issuance are two perpetuating tasks, they cannot be stopped.

The present attempt for introducing ID card is really praise-worthy. But, if the whole affair is not considered under the perspective of our country, it is not unlikely that the 1995 situation will be repeated. I would like to draw attention to the following points.

Incorporation of biometric information: In order to prove that the card-holder is really the person

whose information is written on the ID card, it is not sufficient to include only the bio-data and photograph in the ID card. There is no alternative to inclusion of biometric information if one wishes to stop the production of fake ID cards. The card will also not be accepted at the international level if we do not follow the standards stated in "Machine Readable Travel Documents (MRTD)," prepared by the International Civil Aviation Organization (ICAO). The ID card of Malaysia, under the name MyCard, is the most advanced ID card in the world. A micro-chip having a capacity of 8 kilobytes of memory is the integral part of the ID card. The bio-data, photograph and finger-print of the card holder are stored in the chip. The multipurpose use of MyCard has made life easy. On the other hand, Canada has included facial information in its national database. Facial biometric information is not simply photograph; rather, several features of the face are stored as biometric information.

Whatever biometric feature is used for the identification of a person, it takes a long time to search a person from the database. It becomes time consuming when the size of the database is huge. If we want to develop a system where a person can be searched within an acceptable time we have to use "multiple biometric information" because we have a population of 150 million. Again, the equipment available to identify a person on the basis of biometric information are mostly manufacturer-dependent. If we collect data using a particular machine specification, that data cannot be used by machines of another manufacturer. The other

point is that the use of identification software (e.g. AFIS) should also be considered. If we do not keep the data compatible with such software, they may not be effectively useable in the future. Has anyone spent sufficient time in order to finalize these issues before placing any proposal?

The proof of nationality: During the registration process the nationality of a person must be confirmed so that a foreign national does not get an ID card. At present, the process followed during the issuance of passport is time-consuming and also full of corruption. The police department does the job. Under the present circumstances, a panel of renowned personalities of a locality may be formed to execute the job of identifying a person, whether he is Bangladeshi national or not.

Designing the forms for data collection: In fact, the success of the whole project depends on the design of the data collection system. Normally, the data is transferred from a normal or computerized form after it is collected. First, it is required to fix the data digitization process. It may be by manual data entry, or by using OMR form or OCR form, or also by sending the data from the field through a mobile network after direct entry using a laptop. There are two techniques of manual data entry. In one technique the data entry operator directly looks at the paper form laid beside his computer and enters the data. In the other technique, the data entry operator looks at a scanned image of the data form in his PC monitor and enters data. There are several relative advantages and disadvantages of all the methods of data digitization.

We, at IICT of BUET used a data entry system using OMR forms in SSC/HSC Examination Computerization Project. Again, in the project of Technical Board Examination System we used OCRbased data entry system, and in Bakhrabad and Jalalabad Gas computerisation system scanned-image assisted data entry system has been used. In between 1990 and 2002, normal data entry system has been followed for Rohinga refugees under UNHCR, and for Desa and T&T billing etc. Currently, we are using mobile phone network to send on-line data from power systems. On the basis of the experience of using different options, and considering the qualification and efficiency of data collection teams of our country, we are of the opinion that the use of OMR or OCR -form based data entry will not be fruitful. On the other hand, the use of laptops is slow and costly. Under these circumstances, use of conventional manual data entry paper forms will be the logical choice for our country. But, the data collection form should be designed such that it facilitates fast, error free and easy manual data entry.

Collection of data: The qualification and expertise in running the equipment/computer are the points for consideration in designing data collection process. The data collection team should be properly trained. Mere assumption is not sufficient that this specialised job can be executed by any disciplined team. We already have teams in the field level that can collect hand-written information. The Election Commission may be able to mobilise such a large team within weeks. They comprise mainly of school teachers. On the other hand,

how many of them will be able to fill up a computerised form smoothly and run the equipment at the same time is the question. After proper training, and after running a pilot project and modifying the system design (if necessary), we should be able to predict the time requirement properly.

The calculation should not be based on just following mathematics of multiplication. If it is a matter of collection of bio-data only, I would say a more or less correct prediction can be made. But, if the data collectors have to operate a camera for taking the photographs, run a laptop to synchronise the various data, operate a finger print scanner to capture the finger print, and manage a mobile generator to supply the power, I am in favour of a cautious time prediction.

Data digitisation: Since a large volume of data has to be digitised by entering data from the forms, it will take a long time if the data entry is to be made at a central location. The alternative is to establish data entry centres at the districtlevel. We should not go beyond that level from the management point of view. Even when we think of district data entry centres, it is not easy to create 64 district centres overnight. It may be said that there are a large number of IT firms; each district centre may be assigned to a firm. Yes, that sort of idea has to be implemented; there is no point in keeping this part of the job at the government level. The point is that the IT firms alone cannot do the job by themselves. They have to appoint data entry operators, data control supervisors, centre chiefs etc. Since we do not have ready-made people, we have to develop those people. We have to give sufficient time for that.

Information integration: The proper integration of bio-data, photograph and biometric information is a very sensitive issue. Since the digitization of the bio-data and that of photograph and biometric data are made separately, the accuracy of proper integration will depend on the system followed during collection and integration processes. In case of integration, a common tag should be used. If the common tag is not automatically readable, there are chances of errors during integration of different types of information. In this case, bar code may be used. If the same bar codes are used for the digitization process of bio-data, photograph and biometric information, it is unlikely that they will be mixed up.

Identifying duplicate registrant: Using the biometric information, duplicate entry should be found out so that the same person cannot register from different places. All data should be collected at a central place using a network, and the duplicate data should be discarded after identifying them by software. If we do not use very powerful multidimensional biometric identification software, the process of duplicate screening will take a huge time. So, this issue must be considered during the system design.

ID card printing: One cannot use ordinary printers for printing ID cards having bio-data, photograph and biometric information. It needs special ID card printers. Now-a-days, very sophisticated ID card printers are available. In these printers blank cards are inserted from one side, in stages the bio-data, photograph, biometric information are inserted on the card and they come out of the printer after fine lamination and verification process. We used such a printer for the smart card that we used in our prepaid electricity meter. Of course, considering the volume of ID cards a number of printers having higher speed will be needed.

Central server and back-up system: We have to use very robust machines for storing the data in our national database. We cannot afford a single point of failure. So, redundancy is a must. Moreover, the infrastructure should be made in such a way that a

fibre optic connected remote back-up server maintains a constant mirror and comes into operation instantly in case of failure of the main server. Since very sensitive data will be stored in the servers, data security must be ensured at the utmost level so that no hackers can breakit.

We shall welcome the present government's initiatives of the ID card project if the above points receive due consideration. Although the previous government declared ICT sectors as thirst sectors, they kept ICT initiatives in their files. Because of this, until today, neither an acceptable ICT infrastructure has been created nor has the required manpower been developed. We notice a large number of foreigners in the top-levels of our software/IT firms. No consolidated effort has been made so far to develop this sector under a well-thought plan and monitoring process. We have to appreciate our practical situation. We should develop our systems on our own; otherwise we cannot progress in this sector. Whenever the data of a citizen is required different agencies create their databases on their own.

Computerised birth registration process has been started, vehicle registration process is on-going, although weak, but the passport department has a database, police department maintains its own database, board of revenue manages a separate one, special branch again operates a database at the land and sea ports. The ideal situation would be to develop a central national database, having key information of every citizen, and different agencies will keep only the additional information of their need on the basis of the national ID. Today or tomorrow we have to establish national database. ID card should not be our target; it can be easily generated if we have a strong national database.

The writer is Professor and Director, Institute of Information and Communication Technology, BUET.