2014-10-12 Although Latex does a great job in justifying and hyphenating normal text, it fails badly for URLs, especially in bibliogra- phies. Further more, I believe that Bibtex is the wrong approach. It re- quires references to be classified as one of a predefined set of bibliographic classes. Any new class, like websites, or any unusual class (as telephone interviews) leads to problems. They are circumvented by putting the values in those attributes that get printed and not in those that would semantically match. I think it's bad that Bibtex has those predefined classes and that it doesn't simply print all available fields, as they would prob- ably add value. (If I specify an ISBN it should get printed in any case.) In my eyes, troff's refer(1) does a much better job on the task. There I don't need to classify my references at all. I just fill in those attributes that make sense and refer will choose an appropriate display template, ensuring that all given attributes are printed. This is how it should be done. Gmane has not yet add the mmh mailing list to their archive, hence I haven't written my welcome mail yet, but Philipp has al- ready sent two patches to the list. ;-) http://marmaro.de/lue/ markus schnalke