Over the past few months, but mostly in the last month or so, a project has been running to convert all of the ODP to UTF-8 character encoding, from the myriad of different (ISO-8859, KO-18R, Shift-JIS, ISO 2022, and so on) encodings previously in use.
This project is progressing well. All of the /World categories are converted (might be one or two small ones missed perhaps?), but in some places there is now a duplicate category. This occurs if the category has accented characters, or non-Latin characters, in the Category name. This will leave behind one old category with the old ISO-8859 (or whatever) encoded name, and page of sites, and the new UTF-8 encoded category name, with it's page of UTF-8 encoded data. An automatic process is going through and taking the old pages out. Additionally the relcat and altlang links are being updated to reflect the new category names (well, not new names, but new encoding of the characters that make the name), and errors could be found there too while the job is still "in progress".
This should all be cleared up over the next few weeks or so. Additionally be advised that the RDF dump might have some UTF-8 errors in it next week, if any site entries got corrupted in the changeover. Things are looking good at the moment, but the job still isn't quite finished.
You might find it easier to browse the Google version of the ODP data in some categories (even though that data is a couple of months old now).
<edit>Both of your links worked OK for me at this time. Might have been a temporary glitch, but do bear in mind the changeover of encodings if you have any further problems.</edit>
Thumbshots was working when I posted that message. I didn't realize that they were *that* current. 