There appears to be a problem when copying non-English characters from PDF files on Windows that causes the wrong characters to be stored in the system clipboard and can adversely affect the metadata creation process. We are investigating the root of the problem and will update when more is known.

In the meantime, we recommend that contributors who create metadata using the adminstration tools should copy-and-paste titles, author names, and abstracts from original MS Word (or other word-processing) documents rather that from the final PDF files (though the actual, final PDFs should be uploaded).