Hi, On Mon, Feb 14, 2011 at 1:39 PM, Karen Coyle <li...@kcoyle.net> wrote: > The MARC record structure would allow the use of more than one character for > the subfield code, e.g. "aa" instead of just "a". (Up to 9, BTW, since it's > a one byte numeric field). That would give us scads more possibilities, but > would require a lot of coding changes for software that processes MARC > records. The number of characters in the subfield code is encoded in the > leader, so we could actually mix 1-char and 2-char records in a single > dataset, but most code that reads and writes MARC doesn't use that Leader > byte to control the number of characters -- we tend to assume "1".
Indeed, actually requiring that MARC parsers respect the Leader/11 would break many of them, and I suspect a fair amount of ILSs have the assumption that a subfield code is one octet wide baked in higher up the application stack. At least as far as most software is concerned, permitting upper case subfield codes would be easier to handle. Out of curiosity, does anybody know if there are (or were) any ISO2709-based formats that ever set the subfield label length to any value other than 2 or set the Leader/10, Leader/20, Leader/21, and Leader/22 to values other than '2', '4', '5', or '0', respectively? Regards, Galen -- Galen Charlton gmcha...@gmail.com
