_pre_000x_sep_000x_sep_000x_post_ format: upper [from_char[,to_char]] makes field UPPERCASE default: whole field upper_w [word_nr[,...] make words in field UPPERCASE default: first word substr from[,len] returns substring from some character in some length default: len - rest of the string skip2nr skip all alphanumeric characters and return just numbers after it. To output "1992." from "cop. 1992." initial [word_nr[,...]] make word(s) into initials (upper case first char and append dot after it) default: first word filter: name of filter which is stored in filters/[name].pm Fields Fields which are read from isis are in format 999x where 999 is three digit ISIS field and x is isis subfield. There are three special ISIS fields supported: 10 and 11 (used for ISBN and ISSN, those are only fields which have just two digits) and MFN which returns ISIS MFN. Formatting fields It's done using simple algorithm: it first selects prefix and then inserts fields (and their separators) or just skips them (so it groups fields and separator right from it). At end it just appends suffix.