Kista, Stockholm, Sweden:
A systems trial within the European Union financed Cybervote project was recently carried out in Kista, Stockholm, Sweden. This trial concerned Internet voting. It invited participation from the elderly citizens in Kista. Much work turned out to be needed to attract and motivate mostly technically unexperienced voters of age 55+ to take their time to participate. The trial was open all day the full week Jan 27 to 31, 2003. At the end of that period, 226 voters had participated.
There were three kiosk areas organized for the voting, in or nearby public libraries, with possibilities for the voters to discuss the topics to be voted on - before and after voting. Coffee was served, with cakes .... This deliberation phase was considered to be of importance.
One could vote from home, but this was only tried purely technically.
The voting turned out to be spread fairly evenly over the week. Almost every voter needed some personal explanatory support, normally over a quarter of an hour, while voting. Many voters had never touched a computer before, and were not familiar with the function of a mouse. After voting, each voter was asked to fill in a thorough evaluation enquete, and there are now important usage data available, data that are being analyzed at the present moment.
Questions to be voted on were decided in an earlier and smaller preparatory test, where the same technology was used. The questions therefore were of evident relevance for the voters.
Voters were asked to preregister, for security. These people were provided with a pin code, in an early surface mail letter, to be used at secure identification. It was also possible to vote without preregistration, by showing up with some secure identification (driver�s license, passport). Each such late voter was provided with a "dummy" name and pin code. This was done only at one of the three kiosks, in order to make sure that each voter could only vote once.
One could vote on PC or on a mobile unit - the Nokia Communicator. Only few voters chose to use this mobile unit, however. During the full trial, there were few problems with the technology. The Internet connected server was situated in an office nearby. It was used a separate server part for the Communicator data, and - for security - it was chosen not to merge the two data files at the final tallying.
The tallying at the end of the trial went without major problems. (A few trials was needed). It was chosen three separate talliers, who needed to cooperate at the tallying. It took only a quarter of an hour to carry through the tallying. The result of the trial (without the mobile votes, that was considered to be more of technical relevance) is shown below:
1 What do you prefer for the future of the Jarvafaltet area?
1.1 To keep the area green 160 voters
2 What type of public communication do you prefer:
1.2 Industrial expansion 17 voters
1.3 Expanded living possibilities 28 voters
1.4 No answer 6 voters
2.1 Subway 50 voters
3 What cultural expansion do you prefer?
2.2 Tram 103 voters
2.3 Buses 44 voters
2.4 No answer 14 voters
3.1. A cultural building (Kultur-hus) 114 voters
As a summary, underlining the work needed to support each senior Kista voter during the voting process, the participatory and technological outcome of this trial - testing new means for increased local deliberative democracy - can be considered successful.
3.2 An outdoor theater 43 voters
3.3 A stone age village 25 voters
3.4 No answer 29 voters
Kista trial on the web :
Issy les Moulineaux, France:
860 volunters electors have elected, on Wednesday December 11th, their representatives to the city borroughs' counsels, using the CYBERVOTE system. The voters could vote either from home or from oneof the traditionnal polling station. This election was only an electronic voting one, and no paper voting was carried out in parallell
This election had been approved by the CNIL (French Data Protection Office)
Some problems were encountered with people voting from home with a modem : it concerns the duration of the voting process (the download time could reach 45 minutes before being able to vote) : This problem was due to network overload and the big size of the cybervote applet.
Via ADSL or cable connection, no difficulties appeared.
Concerning citizens voting from their company, as most of the companies were equipped with broadband Internet access, the voting process was easy done. Nevertheless a main problem have been identified: The Cybervote application, with the Issy configuration, uses a special communication port. For security reasons the networks of most of the companies, do not permit linking with this special port. As consequence voting was not possible for some people in their companies.
For the majority of people (80%) having voted from a polling station, no problems were encoutered.
The count of votes took place at 9.00 pm at Issy-les-Moulineaux's City Hall with the attendance of the CNIL, the Ministry of Interior, the Ministry of Industry and CNRS representatives and specialists for Internet voting. The count was done sucessfully in less than ten minutes.
Issy-les-Moulineaux trial on the web :
The German online elections realised within the framework of the CyberVote project were carried out at the Bremen University for Public Administration on 13-15 January 2003.The trial covered the elections of the three University's representative bodies: the university council, the councils of the different university departments and the student council. In addition to traditional ways of voting, voters holding a digital signature card (smart card) could cast their vote online.
In a federal political system like Germany it often happens that elections are combined in one poll. In this case the voter has to cast several votes within a single poll because he has to elect different regional administrative bodies at the same time.
The elections at the University of Public Administration offered the possibility to simulate this kind of political online elections as these elections combine the election of one single candidate and list elections as well as two elections in one polling (elections for the university council and the faculties councils).
In the forerun of the trial an adaption of the legal framework was necessary. As postal vote was already allowed, it was possible to tie up to the prevailing case law in Germany concerning postal vote and to add the option of online-voting in the University's electoral regulations. In order to reach the critical mass of participants, a student's project was organised at the University initiating a group of students in the process of online-voting and trial preparations.
These students acted as multipliers with the objective to motivate and sensitise potential voters at the University. Technical assistance was given to support potential voters when applying and activating their smart cards.
For the setup of the trial elections the certificates of the smart card holders had to be registered for the Cybervote system. This was due to the fact, that in the Bremen trail we did not use the possibility to derive voters public-private key pairs from a voter chosen password, but instead we used the already registered public key being part of the certificate stored on the smart card.
These certificates are PKI based and issued by German Telecom's trustcenter TeleSec.
However, using PKI-certificates without any further identifying information besides certificate holder's family name, first name, a name identifyer (for name doubles) and a public key is insufficient for identifying a voter online. Thus, a registration step prior to the online election was introduced were the voters had to be present.
During this registration, which just took about 30 sec., voters had to insert their smart card into a card reader and the certificate was read from the card. In a subsequent step the public key was extracted from the certificate, stored in the Cybervote system and assigned to the respective record in the voter database. Here, also the assignment of voters to the different voter groups (professors, staff, student) was done, as this information is as well not stored in a standard X.509 certificate.
For the voting act a PC with a smart card reader connected was installed in a corner of the library of the University for Public Administration. Thus, both aspects 1) privacy of voting and 2) protection of the hardware against vandalism were covered. Access to this polling station was given during the opening hours of the library, i.e. from 8:00 a.m. to 5:00 p.m. on three days from Monday, Jan. 13th. until Wednesday, Jan. 15th.
The voting server was set up in the server room of the Statistisches Landesamt and secure client-server communication was realized within the intranet domain of the City of Bremen.
During the period of the Online Elections a totel of 47 voters cast their vote(s) online. The low number of voters was considered disappointing compared to the effort spent to equip people with smart cards (only about 1/5 of those who applied for a PKI smart card used the card in the poll, which gives indication about acceptance problems with this type of cards).
However, the basic approach of using PKI based smart card in combination with asymmetric encryption technologies is still considered to be the right track to secure online voting. What is missing are enhanced certificates with additional information (i.e. date of birth, place of birth) that would allow together with first and family name a real identification of a person in online transactions.
The actual voting only took about 60 to 90 seconds including authentication (by entering the PIN of the smart card on the card reader num pad), marking the candidates on two (for students three) ballots and removing the smart card. In this respect, all technical and organisational requirements as previously defined by the end user partner Bremen were met.
The only bitter taste was the low participation in this online election what, by the way, reflects the general attitude of political awareness in this social environment: the overall participation (including paper ballot elections) was below 10% !
In the afternoon of the third day the Online Voting was closed and in presence of three previously assigned talliers (members of the three different voter groups) the election result was calculated. Again, the process was very fast. The tally was calculated within 10 minutes and the election result (of the online election) was announced.
An important feature of the Cybervote system could be deployed in this case: as only one single voter of the voter group "staff" cast his vote online, the encrypted vote was removed from the ballot prior to tallying and the voter was informed that he has to participate in the paper ballot elections (which started only after the Online Election was closed).
The overall conclusion from the end user Bremen point of view is, that the Cybervote system, as developed prototypically, works well and in compliance to the predefined technical, organisational and legal requirements. Furthermore, the use of a PKI in combination with smart cards is considered to be the only way to accomplish both strong authentication and secure transaction in eGovernment.
However, the trial elections together with other observations made in the field of eGovernment developments reveals severe acceptance problems at the users' side. There is still a certain level of mistrust in this new filed of application.
Although people are using smart cards for bank transactions as a matter of course (and thus relying on the same technology used here) they are still reluctant when it comes to elections and a political vote to be cast in an environment that is not as easily comprehensible as a polling booth.
Bremen trial on the web :
Back to top