With the advent of the Linked Data initiative and with the increasing speed of globalization it becomes urgent to link the unconnected data sets in possession of public and private institutions. The last ten years saw an expanding effort for connecting several data sets to the world's biggest encyclopedia, Wikipedia (e.g., YAGO). This linking effort it is not new, but it is part of continuing work to interlink lexical ontologies like Wordnet to more formal ones like SUMO or CYC. In this talk, we will discuss the techniques that were used for automatically connected lexical ontologies such as Wordnet to Wikipedia or link entities in text documents to corresponding Wikipedia concepts. Moreover, we will present methods to evaluate the accuracy of the linking process. Knowing the linking methods and their pitfalls will enable data custodians to choose the right tools and techniques for connecting heterogeneous data sets to Wikipedia.