Really well-made and easy to use (I'd just change the interface font to proportional). OCR text seems to be without errors beyond the occasional "hyphe- nation." Photos are downloadable in high quality, and they also link to full browseable issues. The dataset cuts off in 1963 though, hopefully it'll expand as more material becomes available in the public domain.<p>Edit: found some examples of bad OCR too. For example, half of the few results for "Taiwan" are misspellings of another word (the words are actually "Fagan," "Edward," and "Japan").