Dear All, I am currently engage in some collaborative research looking at minority language use on YouTube and this has raised some methodological and practical issues that I was thinking AoIR folk might have a view on. Firstly there appears to be no reliable way of identifying videos in minority languages (other than by watching them!) - the use of the name of the language as a search term seems unreliable and unrealistic in terms of probable user behaviour (our minority language Welsh speaking users are fluent majority language English speakers too). Ideally it would have been nice to search for the type of material that users are interested in and/or would like in the minority language - but determining that would be a whole other piece of research. As a slightly better (possibly!) approach we are intending to use the location based search (though we are still trying to get to the bottom of how location is actually specified and what it actually specifies). This should the provide us with seed videos and we can then follow various links within YouTube to expand our dataset. The practical point we were wondering about is whether anyone had actually managed to track down a contact point in YouTube to discuss the possibility of them providing datasets for research purposes etc? Your views on any of the above would be much appreciated, particularly any clever tricks or pitfalls we might have overlooked - thanks! Daniel. Daniel Cunliffe Computing and Minority Languages Group, University of Glamorgan, Wales, UK. http://www.comp.glam.ac.uk/pages/staff/djcunlif/