David Sklar writes about government funded open source at O'Reillynet. This time, the seed is the German Government's funding of Kroupware, an open source alternative to an all-in-one information manager like Outlook.
But does a government's responsibility go further than just a cost-benefit analysis? When a government is paying for software, that money is (usually) coming from its citizens. Should those citizens expect to get something more for their dollars or Deutschmarks than just the ability of government administrators to process words and crunch numbers? They should expect the government-funded software to be available for public use under open-source licensing schemes.Public funding of Open Source is a very delicate issue. In many cases, David's logic does make sense; does HUD, Parliment or the Bundestag need to pay exhorbitant licensing fees for a webserver when Apache is best-of-breed, industry-standard, open source replacement.
In other cases, this common argument doesn't make sense. Where is the best-of-breed open source screen reader for blind users? Where's the industry standard open source Enterprise Resource Planning Tool? Where is the open source HR management application?
Perhaps there's a reason for a government to start a consortium to produce open source enterprise applications. Two issues must be resolved before that can happen. First, most of the domain knowledge needed to write these kinds of niche applications. Second, there is a very strong disincentive for those with this domain knowledge to put it into the public domain, through the development of open source software.
To date, open source has succeeded where the domain knowledge is widespread: compilers and operating systems have been studied and implemented for about 40 years; databases have been studied and used for about 30 years; windowing systems have been studied and implemented for about 20 years. How long have ERP systems been commercially viable? Off-the-shelf HR systems? Screen readers? Speech-to-text systems?
