News

A good choice

October 28 was a key day for the Provincial Computing and Communications Department of the Government of Mendoza, Argentina: it was the moment of truth.

This was the day of national and provincial elections in the Republic of Argentina, and the government of Mendoza had to provide query results to both the general public and to the local, national and international media. At the same time, it was necessary to give clear and specific information to citizens about the electoral system, i.e., electoral register queries.

On previous occasions, and for many years, a solution developed in Clipper (DOS) was used for election data loading and processing. Data dissemination entailed the generation of files deployed through a web query written in PHP, without the existence of a portal for population information. The need to accelerate the election computerization process, to improve the service rendered to society, and to take full advantage of the information obtained became evident.

Choosing GeneXus as a technological axis for future solution was easy. The government of Mendoza has been using this type of technology for some years, taking advantage of ACP Ingeniería de Sistemas SRL experience in technology management. The new solution needed to manage the positioning and refining process of a Unified Electoral System, fully based on a distributed platform – the Web- and the mid-term goal of decentralizing electoral data processing, moving towards computerizing the elections 100% long term (internationally called the “electronic vote”).

In order to meet all these requirements, it was decided to look for a solution using several GeneXus technologies. On the one hand, a web portal, featuring an array of contents with public and general information about election processing. This portal, developed with GXportal 4.2 – and executed in a Windows 2003 Server, with SqlServer 2000 as DBMS and Intel Xeon architecture – was the place where all contents associated with the elections were published.

At the same time, a Web-Java system was developed with GeneXus 9.0, which was used for loading telegrams from the ballot–counting tables and for data processing. This system was deployed in a Linux dedicated server (isolated from external access for safety purposes, as data were loaded from Government House) on a Tomcat 5.5 servlet engine.

Finally, an application for data dissemination was developed, deployed in two servers (Linux with Tomcat 5.5) and integrated into the portal, displaying the election results according to the voted posts (President, Governor, Mayor, etc). Information could be viewed in a bar or pie chart by using GXchart 3.0.

Data and solutions used in the past were analyzed, and the circuit of information and the format of forms used during the elections were reviewed, serving as a basis for the design of a system primarily focused on simplifying and accelerating telegram loading. Further, for the purpose of improving response times to public queries, partial data were accumulated by means of an automatic process.

However, the ultimate test for the solution was Election Day itself. As there was no previous Web system, the amount of requests to be received by the pages during the elections could not be predicted. In addition, the fact of using a web system - and that the information involved was critical - demanded higher safety measures for system access and data protection. For this purpose, the system was isolated within an intranet, accessible only from those workstations used for data loading, and a multiple key schedule, implemented for information initialization.

The result couldn't be any better. According to Arturo Caridad, Computing and Communications Director, Mendoza Province, “it was possible to handle the loading and processing of poll data with no problem whatsoever, taking advantage of the simplicity of the installation. Also, the fact of knowing and being familiar with the implementation platform enabled an enterprise setup with a group of software and hardware resources successfully supporting the large number of queries received.”