Linux in Government
Linux has been making serious inroads in government as well as in other markets. We all know why this is. The quick summary is it works well and the price is right.
As in most of the world, Linux has made its presence known in Costa Rica. My interview with Guy de Téramond, which appeared in the January issue of Linux Journal, detailed one example of Linux at work in Costa Rica. It also is running on servers at the Casa Presidential (the Costa Rican equivalent of the White House) and CIPET, a branch of the Ministry of Education that provides technical training for teachers.
I recently met with some Costa Rican system administrators who work in the government. In that meeting we talked about three issues:
technical issues of installing and running Linux;
the politics of getting users to accept Linux on their desktops; and
the upper level politics of getting Linux installed on desktops.
Before I get too far into the specifics of our discussion, let me give you a new resource Linux Journal has made available to help people involved in the "Linux in Government" effort talk to each other. We have a new mailing list, called gov-list. To join the list, go to this URL and sign up. We encourage the discussion of any issues you come across while getting Linux to do what you need within government.
The technical issues seem like the easiest piece of the puzzle. Virtually no difference exists between making Linux work in a government environment versus any other environment. Linux works well; Linux interoperates with other systems; and Linux is reliable. Some unique issues may turn up in the government sector, but I sure haven't seen them.
One item that initially came to mind was the need to prove Linux works without buying new equipment. In fact, Linux can be loaded on existing PC hardware and should perform at least as well as what was previously installed. For hardware with limited capabilities, running Linux-based thin clients is an alternative. With a capable server, slower hardware with limited RAM can perform much better running Linux than it did when running another OS locally on the hardware.
Thus, the problem is really a political problem. Linux offers options, and it is up to you to figure out which options offer acceptable performance.
User politics is one issue that comes up regularly. The good news is Microsoft has done a lot to make this one into a non-issue. Many office environments considering an upgrade currently run Windows 9x. A Microsoft-based upgrade means a change in what the desktops look like anyway. Odds are user resistance will occur whether the transition is to Windows XP or Linux. The fact that there will be a change in the user interface anyway could be a way of slipping Linux in the door.
In many environments, users are not particularly concerned with the OS behind the scenes as long as their applications look familiar. Both StarOffice and OpenOffice, for example, have managed to offer a familiar user interface to keep users comfortable. Similarly, a different web browser, whether it be Opera, Mozilla or Konqueror, performs the same functions as Internet Explorer and can be sold on its features.
If the office environment is composed of different departments, take advantage of Linux's ability to interoperate, and convert a department at a time. Done properly, the first converted department can help promote the overall conversion. In the end, the advantages of increased speed and reliability help offset any differences.
Politics on the upper level is going to be the hardest issue to deal with in many conversions. On this level, politics is likely to take the form of vendor pressure to go with what they consider standard. Many vendors see the use of less expensive software as a loss of revenue, but this does not have to be the case. If an organization has $100,000 to spend, for example, show the vendor that the total expenditure is fixed. Work with them to develop a plan where they acquire happy customers, and they should go along.
Say you have $2,000/desktop to spend. By decreasing the software cost per desktop, more of the money could be spent on hardware. That could mean LCD displays instead of CRTs, faster computers or more printers. All of these items increase productivity and generally create a happier user base.
Take advantage of this model to show the vendor that more business is out there. If they embrace Linux and, as a result, make your agency run better than the next agency, they have a better chance of winning a bid from another agency.
Finally, one other card to play is the "buy local" card. With Linux you have a non-proprietary solution, which means you can support the solution locally. I know, for instance, that the Caja (the Costa Rican equivalent of the US's Social Security Administration) spends over $1,000,000 per year on support from Microsoft. It would be to Costa Rica's advantage to invest that money in local technical support rather than export the money to the US.