I'd like to be able to index the metada without diacritics, so that it can then be found without having to specify them.
I've read a lot about filters, and flattening, but the only patch that worked is the one posted by Antonio: ISOLatin1AccentFilter. It works but only one way, that is, it will not work if the metadata has diacritics inside, because it will be indexed with them and it will only be found if I type it exactly.
user --> ISOLatin1AccentFilter --> DSpace --> ascii metadata (works)
user --> ISOLatin1AccentFilter --> DSpace --> latin1 metadata (does not work)
I've also noticed that that question shows up regularly in the mailing lists so, if there is an answer, maybe it could be added to the FAQ ...