Kyle Banerjee wrote:
> I think it would be worthwhile discussing what you can/cannot count on as
> well as the reliability of certain data points.
> There is a natural tendency among noncatalogers to rely on specifications
> for figuring out what data will be where. However, some fields that may
> seem perfect for an application may not be used at all, certain types of
> data might appear in any one of multiple fields, and other types might not
> be used as expected. Incorrect assumptions about what metadata contains
> and its reliability will lead to trouble every time.
> It might also be worth making some broad remarks on what you can reasonably
> expect to accomplish with records encoded to certain major major standards
> as well as their limitations and what you lose when you crosswalk from one
> standard to another.


Along similar lines: Really knowledgeable catalogers who are familiar with what's happened in cataloging over the past few decades seem (at least, to coders) to be very good at interpreting records that vary from current standards/practices; their knowledge of past standards, past practices, past events, and even regional differences gives them a good basis for interpretation and a well-developed intuition. Even if they don't know off hand why a particular piece of data is expressed a particular way in a particular record, they often have an idea about where to look. Non-catalogers, on the other hand, are generally just flummoxed by data that doesn't appear to follow the spec. Without the requisite background knowledge, there's no way for us to tell cataloging mistakes from meaningful non-spec data, and there's no way for us to intuit the meaning of non-spec data.

So, I'd be interested in hearing: what are some of the canonical events in recent cataloging history that help explain the most common patterns we're likely to run across in the wild? "Events" might refer to a number of things. Maybe official changes in standards; maybe widespread changes in practice; maybe a vendor screwing up a big-time data load and affecting millions of records in some nonstandard way, where catalogers had no choice but just to deal with it. "Patterns" might also refer to a number of things. In addition to actual patterns we might find in the data, maybe sets of standards that tend to go hand-in-hand, like MARC21/AACR2, and maybe common deviations from these standards. What are the versions of AACR2 and what are the telltale signs of each version that we'd find in the data? What other cataloging standards are commonly used with MARC21 and what do they look like? How do we know them when we see them? How have catalogers been instructed to handle data that is non-standard or follows an out-of-date standard when they come across it, and how does that factor into what we end up seeing in records?