I think that in future many people will ask for list of open datasets: by country, by argument (science, geography, art,..). Is a good idea create wiki like these? I think that this wiki could be used also to add new open datasets that you heard of. Or should we start from a question like: "Is there any dataset for Italy?" and convert it later to wiki?
I'm not sure creating a catch-all wiki for this type of misinformation is going to be good for this site.
Remember, the lifeblood of this site is search; Google Search — and if someone is searching for a "Database of politicians for Italy," I sure would like them to find this site.
In the bigger picture, this site is supposed to be about answering specific questions you encounter in your day to day work. This site isn't supposed to be a list of where to find things on the Internet. If you distilled this subject down to a few generic, catch-all lists of where to find stuff, I'm afraid you would miss out on a lot of fascinating problems people are trying to solve.
I would stick to specific questions. Let Open Data SE become a huge repository of fantastically-detailed, long-tailed questions that would serve to separate you from any ol' generic site on the subject. That's how we're going to hold the interest of the experts you are trying attract.
I like this idea, but maintaining such a list might be problematic and as others point out, it's a better idea to empower search engines like Google to find those datasets.
There is a new dataset type added to Schema.org and built on the W3C's DCAT work described at http://schema.org/Dataset The use of this is likely to provide better generic search to actual datasets. A detailed description behind this is at http://www.w3.org/wiki/WebSchemas/Datasets
There is a listing of government open data sites at http://www.data.gov/opendatasites
(Disclaimer: I serve as the Evangelist for Data.gov)