Comment: OpenOffice's Tale of Two Cities

By Oliver Diedrich

Failure in Freiburg, success in Munich. Experiences with open source software in the public sector couldn't be more different. If there's a lesson to be drawn from this, it's "go the whole hog or not at all".

At first sight it looks pretty straightforward – a licence for Microsoft Office Professional 2010 costs just under €400. Add that up over 10,000 workplaces (as is the case in Munich's city administration) and it comes to more than €4 million. For open source alternatives OpenOffice and LibreOffice, by contrast, licensing costs are zero, so you've saved at least €4 million. In view of the state of public finances, you'd think that would be the end of the discussion.

But it's not. Companies and other organisations that buy large numbers of licences do, of course, enjoy significantly lower prices, benefiting from Microsoft's volume licensing programme. Specific figures are hard to come by, but if you're paying three figure prices for five figure volumes of Microsoft Office, you probably need to work on your negotiating skills. Licensing costs also need to be seen in relation to overall costs for each workplace. If an administrator earns €25 per hour, just a few hours of lost productivity per year can quickly negate any savings on licensing costs.

Consequently, open source software in the public sector has never just been about saving money – it's also about reducing dependency on Microsoft. The promise of better integration draws many users ever deeper into the comfortable Microsoft world. Outlook is included with Office and Outlook offers optimum compatibility with Exchange. In turn, Exchange requires a Windows Server with Active Directory and IIS is thrown in for good measure, for which the easiest way of creating applications is ASP.net. And so it goes on. When you eventually find yourself faced with the choice of either aligning your IT systems to Microsoft's priorities or undertaking major restructuring, things can start to get extremely expensive. This is the dilemma currently facing Microsoft Small Business Server users, following Microsoft's announcement in the summer that it plans to replace the product with cloud services. And can you really require local citizens to purchase a specific product if they want to send or receive digital documents to or from council offices?

It is little wonder that the councils in both Munich in 2003 and Freiburg in 2007 voted to migrate to open source software. The outcome of the two processes, however, couldn't have been more different. All 15,000 desktop PCs in Munich's city administration now run OpenOffice and 12,000 of these have switched from Windows to Linux. Millions of euros have been saved and the city is proud of its pioneering LiMux project. In Freiburg, by contrast, after years of torment, the OpenOffice migration project has now been abandoned and the council is reverting to Microsoft Office. A switch from Windows to Linux in Freiburg was never even tried.

It makes you wonder what can be so hard about switching from Microsoft Office to OpenOffice. The question is even more pertinent in council offices, where it's a matter merely of switching from one word processor to another. Much of the council's work involves document templates; simple templates are easily converted or can be rapidly recreated, but converting more complicated templates containing fields, macros and form functions, and converting mail merges from Microsoft Office to OpenOffice, is not quite so simple.

Then there are the notorious specialist administrative procedures: municipalities use up to 100 different applications to perform everyday tasks such as residential registrations, vehicle licensing, business registrations, planning and town clerk activities, youth and social work, and administering nurseries, property and cemeteries. This frequently involves bespoke software written specifically for an individual council's administrative staff. These specialist applications frequently rely on Microsoft Office to generate and print documents, save their data to Access databases or are implemented in their entirety in Microsoft Office (it's amazing what you can do with involved Word or Excel macros). All of this is difficult or impossible to implement in alternative office software.

Both Munich and Freiburg went through this experience. Munich learnt the lessons of its teething troubles and developed WollMux, a tool for administering OpenOffice templates and template text blocks. The tangle of existing IT structures was tamed. The procedure for converting to open source software was modified to incorporate pilot projects in which problems were detected and resolved prior to a wider roll-out. This saved a lot of staff a lot of frustration.

In Freiburg, by contrast, it appears that problems were simply not resolved. Staff were expected to muddle through. In 2012, most desktops were still running the ancient Microsoft Office 2000 in parallel to the almost equally outdated OpenOffice 3.2.1 – some tasks were only able to be performed in Microsoft Office, some required OpenOffice. It is not really surprising that this did little to encourage acceptance of OpenOffice and that it generated a great deal of extra work for the IT department. Against this backdrop, the council's decision to put the OpenOffice experiment out of its misery and allow its staff to get back to working properly is understandable.

If there's a lesson in all this it's that migrating a core software tool such as an office suite doesn't work if it's done half-heartedly. It requires careful planning and has wide-ranging consequences and that means an initial capital cost. In addition, a project of this kind needs to get both users and the IT department on board – in Freiburg it appears to have failed on both counts.