I once went to a meeting sponsored by the EU about how to make data resources for computational linguistics more interoperable. I noted with increasing dismay

I once went to a meeting sponsored by the EU about how to make data resources for computational linguistics more interoperable. I noted with increasing dismay that everything people were saying aimed to check off requirements that met EU definitions of interoperability without doing anything at all to the data. For example, someone asked to produce XML could just submit something like this: <xml> <!-- Please download our dataset from http://example.server/example.csv. --> <dataset type="text/csv" href="http://example.server/example.csv" /> <license>unusably restrictive academic license text goes here</license> </xml> And now I kind of understand that was the goal. The people running that meeting weren't trying to be useful, they were trying to create a bureaucracy that would excuse their lack of usefulness.