Hello Tibor, [...] > I've been digressing a bit, but I hope people may find these titbits > useful.
Sure they are. What I have noticed, though, is that, as this tokenizer is done (of course) during the indexing phase, author browsing is much less useful now to check possibly duplicated forms of the same person, like in those examples: http://invenio-demo.cern.ch/search?ln=en&p=Dasse%2C+Michel&f=author&action_browse=Browse It seems that Darwin or Dasse are written in two different forms, as they appear in full (Darwin, Charles) and in initials (Darwin, C.), but both point to the same records. This had caused us quite a bit of head-scratching during the last few months. Anyway, now we have a clear picture of this author tokenization. Thanks again, Ferran
